在成人的择期手术中食管引流型喉罩气道与经典喉罩气道(laryngeal mask airway, LMA)在正压通气方面的对比
摘要
研究背景
声门上气道装置的发展彻底革新了全身麻醉期间的气道管理。临床实践中广泛使用两种装置来促进正压通气：食管引流型喉罩气道（ProSeal laryngeal mask airway，pLMA）和经典喉罩气道（Classic laryngeal mask airway，cLMA）。目前尚不清楚这些装置在效果和并发症方面是否具有重要的临床差异。
研究目的
比较食管引流型喉罩气道（pLMA）和经典喉罩气道（cLMA）在成人的择期手术中进行正压通气的有效性。
检索策略
我们通过Cochrene图书馆检索了Cochrane对照试验中心注册库（Cochrane Central Register of Controlled Trials, CENTRAL; 2017年，第3期），并检索了MEDLINE（Ovid SP,1997年至2017年4月）、Embase（Ovid SP，1997年至2017年4月）、科学信息研究所（Institute for Scientific Information，ISI）的Web of Science （1946至2017年4月）和护理与健康相关文献累积索引（Cumulative Index to Nursing and Allied Health Literature，CINAHL; EBSCO主办，1982年至2017年4月）。
我们检索了试验注册库以获得截至2017年4月正在进行的研究。
我们未设置语言限制。我们将检索时间限制在1997年至2017年4月，因为pLMA在2000年被引入临床实践。
纳入排除标准
我们纳入了比较pLMA和cLMA对接受择期手术的成人进行正压通气的有效性的随机对照试验（randomized controlled trial，RCT）。我们计划仅纳入与交叉RCT的第一阶段相关的资料。
资料收集与分析
我们使用Cochrane协作网推荐的的标准方法学程序。
主要结果
我们纳入了8项合格的随机对照试验，共涉及829名受试者（pLMA和cLMA组分别有416名和413名受试者）。我们确定了6项等待分类的交叉研究；1项已完成但尚未发表，其他5项研究第一个治疗期的相关资料尚未获得。七项被纳入的研究提供了与主要结局相关的资料，八项研究提供了与多项次要结局相关的资料。
我们的分析受到一个事实的阻碍，即大部分纳入的研究均未报告任何研究组中的事件。没有研究报告两种设备在主要综述结局方面存在的显著差异：未能充分机械通气。我们通过评价两个变量来评价这一结局：氧合不足（风险比 （risk ratio，RR）=0.75，95%置信区间（confidence interval，CI）[0.17,3.31]；4项研究，N=617）和通气不足（不可估计；1项研究，N=80)。
使用pLMA建立有效气道需要更多时间（均差（mean difference，MD）为10.12秒，95% CI [5.04,15.21]；P<0.0001；I²=73%；2项研究，N=434）。cLMA受试者在正压通气期间的气道峰压较低（MD=0.84，95% CI [0.02,1.67]；P=0.04；I²=0%；4项研究，N=259）。pLMA受试者的平均口咽部漏气（oropharyngeal leak，OPL）压较高（MD=6.93，95% CI [4.23,9.62]；P<0.00001；I²=87%；6项研究，N=709）。
根据GRADE评分的评价，所有结局的证据质量均为低质量，这主要与盲法和不精确性相关。
资料显示，在未能插入设备、使用替代设备、粘膜损伤、喉咙疼痛、支气管痉挛、胃胀气、反流、咳嗽和过度漏气方面，两种设备之间没有显著差异。资料不足以估计与设备相关的阻塞差异。没有一项研究将术后恶心和呕吐作为报告结局。
作者结论
由于事件极少，因此我们无法确定这两种气道装置在氧合或通气失败方面的效果。就几种并发症的差异而言，结果不确定。低质量证据表明，食管引流型喉罩气道密封效果更好，因此它可能比经典喉罩气道更适合正压通气。经典喉罩气道的插入速度可能更快，但这不太可能具有临床意义。
PICO
简语概要
哪种气道装置（食管引流型喉罩气道或经典喉罩气道）在成人麻醉期间提供更有效的人工呼吸？
系统综述问题
我们想知道麻醉期间成人使用的一种人工呼吸装置是否比另一种更有效、更安全。
系统综述背景
因外科手术而麻醉的患者通常需要戴上人工呼吸机（呼吸机）以帮助呼吸。这涉及将进入气管或位于声带开口上方（声门上）的装置放入患者口中。我们比较了两种用于促进人工呼吸的声门上装置（经典喉罩气道（the Classic laryngeal mask airway, cLMA）和较新的食管引流型喉罩气道（the ProSeal laryngeal mask airway, pLMA））的有效性。据称，pLMA的优势包括：更好地贴合在口腔内、通气过程中漏气的风险更小以及胃内容物被带入气管的可能性更小。
检索日期
我们检索了1997年至2017年4月期间发表的研究，因为pLMA在2000年被引入临床实践。
研究特征
我们纳入了8项随机研究，共涉及829名18岁及以上的受试者。所有受试者都进行了择期手术并需要全身麻醉。研究人员直接比较了cLMA与pLMA在手术期间提供人工呼吸的效果。
我们确定了6项等待分类的交叉研究；1项已完成但尚未发表，另外5项研究正在收集与第一个治疗期相关的资料，这些资料尚未获取。
研究资金来源
五项研究未报告资金来源。在其余3项研究中，1项研究报告喉罩公司赞助了一些资料，但未参与研究设计、资料分析和手稿准备。另一项研究获得了位于美国加利福尼亚州洛杉矶的Joseph Drown基金会的部分支持。最终研究的其中一位作者获得了两种类型喉罩气道制造商Intavent Ltd的研究资金，但Intavent Ltd没有赞助这项研究。
主要研究结果
我们不确定这些设备在提供充足的氧合作用和通气方面是否表现出了显著差异，因为没有足够的资料使我们得出任何确定的结论。食管引流型喉罩气道可能会提供更好的密封性，因为它在更高的压力下才可能漏气，而经典喉罩气道的插入速度可能更快。然而，这些发现在临床上并不重要。
证据质量
我们将所有纳入的研究都评价为低质量证据，因为麻醉师知道哪些设备被用于哪些受试者（尽管这可能无法避免），并且我们也不清楚收集资料的研究者是否不知道这种干预。这一事实造成了潜在的偏倚。
Authors' conclusions
Summary of findings
| ProSeal versus Classic laryngeal mask airway (LMA) for positive pressure ventilation in adult patients undergoing elective surgery | ||||||
| Patient or population: adult patients (18 years of age and older) for positive pressure ventilation undergoing elective surgery Comparison: Classic laryngeal mask airway (cLMA) | ||||||
| Outcomes | Illustrative comparative risks* (95% CI) | Relative effect | Number of participants (studies) | Quality of the evidence | Comments | |
| Assumed risk | Corresponding risk | |||||
| cLMA | pLMA | |||||
| Inadequate oxygenation (SaO₂ < 90% on peripheral oxygen saturation) | 13 per 1000 | 10 per 1000 (2 to 43) | RR 0.75 (0.17 to 3.31) | 617 (4 studies) | ⊕⊝⊝⊝ | |
| Inadequate ventilation ( ETCO₂ > 45 mm Hg) | See comment | See comment | Not estimable | 80 | See comment | No episodes of inadequate ventilation in either group |
| Time required for effective airway (seconds) | Mean effective airway time in the control group ranged from 22.24 seconds to 31 seconds | Mean effective airway time in the intervention group was 10.12 seconds longer (95% CI 5.04 to 15.21) | 434 | ⊕⊕⊝⊝ Lowb | Outcome added post protocol as some included studies looked at this outcome. | |
| Peak airway pressure during positive pressure ventilation (cm H₂O) | Peak airway pressure in the control group ranged from 17 cm H₂O to 24.5 cm H₂O | Mean peak airway pressure during positive pressure ventilation was higher in the pLMA group by 0.84 cm H₂O (95% CI 0.02 to 1.67) | 259 (4) | ⊕⊕⊝⊝ Lowc | ||
| Oropharyngeal leak pressure (seal pressure) (cm H₂O) | Oropharyngeal leak pressure in the control group ranged from 14.64 cm H₂O to 26 cm H₂O | Mean oropharyngeal leak pressure (seal pressure) was higher in the pLMA group by 7.17 cm H₂O (95% CI 4.85 to 9.5) | 709 | ⊕⊕⊝⊝ | ||
| Sore throat | 194 per 1000 | 150 per 1000 (107 to 210) | RR 0.77 (0.55 to 1.08) | 653 (5) | ⊕⊕⊝⊝ Lowe | |
| Regurgitation | 17 per 1000 | 13 per 1000 (4 to 45) | RR 0.8 (0.24 to 2.69) | 597 (4) | ⊕⊕⊝⊝ Lowf | |
| *The basis for the assumed risk (e.g. median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence. | ||||||
| aThree studies reporting this outcome were assessed as demonstrating adequate allocation concealment and one was unclear. Blinding of participants, personnel and outcome assessor were assessed as unclear in three studies and one study was unblinded. Furthermore, the results are imprecise, as only one out of four studies has reported positive events, therefore evidence is downgraded three levels. bData from two studies with imprecise results resulted in downgrading of evidence by two levels. cBlinding of participants, personnel and outcome assessors were unclear in three studies and one study was unblinded. This resulted in downgrading the evidence by two levels. dBlinding of participant and personnel were unclear in all six studies while blinding of outcome assessor was not conducted in one study and was unclear in the remaining five studies. Data analysis indicated high levels of heterogeneity (inconsistency). For these reasons, the evidence was downgraded two levels. eBlinding of participant and personnel were unclear in four studies and was not done in one study, while blinding of outcome assessor was not conducted in two studies and was unclear in the remaining three studies. This resulted in downgrading the evidence by two levels. fBlinding of participant and personnel were unclear in all four studies while blinding of outcome assessor was not conducted in one study and was unclear in the remaining three studies, therefore evidence was downgraded two levels. | ||||||
Background
Description of the condition
The history of airway management in humans spans almost 4000 years. Modern endotracheal intubation dates back to the 18th century, when obstetricians and first responders first used breathing tubes. It was as late as 1945 when endotracheal intubation was adopted for usual practice (Luckhaupt 1986). Improved understanding of the pathophysiology of the upper airway led to the development of newer, less invasive airway devices such as the laryngeal mask airway (LMA), which has been assigned an important role in anaesthesia practice (Szmuk 2008). In contrast to supraglottic airway devices (laryngeal and pharyngeal airways), the endotracheal tube is a definitive airway device that not only allows delivery of oxygen, anaesthesia, and other drugs, it also enables effective ventilation without increased risk of gastric insufflation and pulmonary aspiration. Although effective, laryngoscopy and endotracheal intubation are associated with morbidities ranging from minor complications such as sore throat and trauma to the larynx, lips, teeth, tongue, and nose to more serious effects such as difficult or failed intubation and noxious autonomic reflexes (hypertension, tachycardia, bradycardia, and arrhythmias), which can be detrimental in some cases (Blanc 1974).
Description of the intervention
The laryngeal mask airway (LMA), designed by Dr Archie Brain in 1988, is a minimally invasive airway device that offers several advantages over conventional laryngoscopy and tracheal intubation. Its benefits include improved haemodynamic stability, reduced anaesthetic requirements, avoidance of muscle relaxation agents, and reduced coughing (Brimacombe 1995) and sore throat (5.8% vs 34% and 14.4% vs 50%, respectively) (McHardy 1999). Since it was introduced, the LMA has gained widespread acceptance over the endotracheal tube as a safe and effective airway device for spontaneously breathing anaesthetized patients (Benumof 1992; Benumof 1996; Bogetz 1994; Jones 1995; Lopez‐Gil 2005; Todesco 1993; Villars 2005; Wat 2003). The Classic LMA is an autoclavable and reusable laryngeal mask airway that consists of an airway tube connected to an inflatable mask with a silicone rim. The single cuff provides a seal around the laryngeal inlet while enabling spontaneous ventilation of the airway and application of positive pressure ventilation (Brain 1983; Brain 1985; Brimacombe 1996; Van Damme 1994; Verghese 1996). However, the seal provided by LMA does not guarantee prevention of pulmonary aspiration, thus revealing its major limitation for positive pressure ventilation (Keller 2004).
Limitations of the Classic LMA (cLMA) (lack of protection from aspiration, airway leak, and risk of gastric distension with positive pressure ventilation) led to development of the ProSeal LMA (pLMA), which has a better anatomical fit and is more suitable for positive pressure ventilation.
The ProSeal laryngeal mask airway (pLMA; Intavent Orthofix, Maidenhead, UK), introduced in the year 2000, was designed by Dr Archie Brain and is based on the same principle that led to the design of the Classic laryngeal mask airway (cLMA). Modifications in design of the pLMA include separation of respiratory and gastrointestinal tracts, provision of a larger and softer wedge‐shaped cuff, and incorporation of an oesophageal drain tube. These modifications were designed to improve safety and efficacy by improving the airway seal, and to enable more effective positive pressure ventilation, reducing the chance of gastric insufflation and consequent regurgitation and aspiration of gastric contents (Brain 2000).
How the intervention might work
The cLMA, in contrast to an endotracheal tube, cannot guarantee protection of the lungs from aspiration of regurgitated material. Deep (subglottic) suctioning cannot be performed through the cLMA. Although the effectiveness of cLMA in preventing pulmonary aspiration has been documented, evidence of gastric distension and pulmonary aspiration has been noted with cLMA use (Griffin 1990; Maltby 2000).
The pLMA can provide higher oropharyngeal leak pressure without an increase in mucosal pressure. This enables more effective positive pressure ventilation without the need for high cuff mucosal pressure (Brimacombe 2002; Keller 2000). It has been suggested that the pLMA double‐cuff design enables twice the leak pressure for a given intracuff pressure compared with the cLMA (Brimacombe 2002). A further potential advantage of pLMA is that the drain tube provides access to the oesophagus, enabling insertion of a gastric tube, which may reduce the risk of aspiration (Evans 2002; Mark 2003).
Why it is important to do this review
Classic LMA is a safe and effective device for use in anaesthetized patients who are breathing spontaneously. A meta‐analysis of cLMA studies reported only 0.02% risk of aspiration, but this analysis did not look specifically at the risk of aspiration with positive pressure ventilation (Brimacombe 1995a).
Classic LMA has been studied extensively for positive pressure ventilation (Bernardini 2009; Devitt 1994) and has shown benefit for select patient groups, including those with normal lung compliance and airway resistance and people for whom airway pressure does not exceed 20 cm H₂O, which causes gastric insufflation. Classic LMA has been shown to be effective as an endotracheal tube for positive pressure ventilation. Other consequences of positive pressure ventilation through the cLMA include hypoventilation and operating room pollution resulting from leakage of inspired gases due to higher seal pressure.
Given that pLMA has a portal for gastric deflation and higher oropharyngeal leak pressure, theoretically it should be a better choice for use in positive pressure ventilation.
Objectives
To compare the effectiveness of the ProSeal laryngeal mask airway (pLMA) and the Classic LMA (cLMA) for positive pressure ventilation in adults undergoing elective surgery.
Methods
Criteria for considering studies for this review
Types of studies
We included randomized controlled trials (RCTs) comparing the effectiveness of positive pressure ventilation by Classic and ProSeal laryngeal mask airways (LMAs) in adult patients undergoing elective surgery. We planned to include data related to the device used first in cross‐over RCTs.
Types of participants
We included adults (18 years of age and older) scheduled for elective surgery who required general anaesthesia with positive pressure ventilation.
Types of interventions
We included studies that compared use of the Classic LMA (cLMA) versus the ProSeal LMA (pLMA) for positive pressure ventilation under general anaesthesia using:
-
any technique of LMA insertion;
-
any type of cLMA (reusable or disposable) and reusable pLMA; and
-
any mode of positive pressure ventilation.
Types of outcome measures
We anticipated a range of definitions for primary and secondary outcome variables among the included studies. We therefore intended to subgroup such outcomes and to consider pooled results, if appropriate data were available.
Primary outcomes
Failure to adequately mechanically ventilate according to two variables
-
Inadequate oxygenation measured as less than 90% on peripheral oxygen saturation (SaO₂)
-
Inadequate ventilation assessed by end‐tidal carbon dioxide concentration (ETCO₂) greater than 45 mm Hg
Secondary outcomes
-
Inability to insert laryngeal mask airway or abandonment of insertion of the intended device ‐ assessed by inability to insert or use an alternate device or both
-
Effective airway time ‐ assessed as time taken to insert and adequately ventilate using either device
-
Failure to insert gastric tube
-
Peak airway pressure during positive pressure ventilation
-
Oropharyngeal leak pressure
-
Complications related to LMA insertion: mucosal injury (blood staining), sore throat, excessive leak, gastric insufflation, bronchospasm/laryngospasm, regurgitation/aspiration, coughing, obstruction of device, excessive secretions, postoperative nausea or vomiting
Search methods for identification of studies
Electronic searches
We searched the Cochrane Central Register of Controlled Trials (CENTRAL; 2017, Issue 3) in the Cochrane Library (Appendix 1); MEDLINE (Ovid SP, 1997 to April 2017) (Appendix 2); Embase (Ovid SP, 1997 to April 2017) (Appendix 3); the Institute for Scientific Information (ISI) Web of Science (1946 to April 2017) (Appendix 4); and the Cumulative Index to Nursing and Allied Health Literature (CINAHL) (EBSCO host, 1982 to April 2017) (Appendix 5).
We imposed no language restrictions. We restricted our searches to the time from 1997 to April 2017 because pLMA was introduced to clinical practice in the year 2000.
Searching other resources
We handsearched reference lists of all included studies for trials missed by the electronic search strategy.
We also searched the following trial registries for ongoing studies in April 2017.
-
http://www.clinicalstudyresults.org (we were able to search this website only to December 2016. Thereafter, the website became unavailable).
-
WHO International Clinical Trials Registry Platform (ICTRP).
Data collection and analysis
Selection of studies
Two review authors (QH, KS) independently assessed titles and abstracts identified by the searches. We obtained full‐text copies of potentially relevant trials for assessment according to parameters outlined in Criteria for considering studies for this review. We assessed trials that met those criteria for methodological quality only.
Data extraction and management
Two review authors (QH, KS) independently extracted data using a data extraction form for each trial (Appendix 6). One review author (KS) entered all data into Review Manager (RevMan 5.3), and a second review author (QH) checked these data for completeness and correctness. If required, we planned to contact primary authors of trials to request information, especially regarding unpublished data.
We planned to extract the following data for each included trial.
-
Participants (age, gender, body mass index (BMI), sample size, allocation, flow of participants throughout enrolment, follow‐up, analysis, protocol deviations if any).
-
Interventions.
-
Technique used for cLMA and pLMA insertion (with or without introducer).
-
Epigastric auscultation.
-
Oropharyngeal leak.
-
Drainage tube leak.
-
Oropharyngeal leak pressure.
-
Confirmation of correct placement with fibre‐optic bronchoscope.
-
Ease of insertion.
-
-
Ventilation parameters: peak airway pressure, inspired tidal volume, expired tidal volume, pulmonary compliance, blood oxygen saturation (SpO₂), partial pressure of oxygen (PaO₂), ETCO₂, partial pressure of carbon dioxide (PaCO₂).
-
Outcomes (primary and secondary).
Assessment of risk of bias in included studies
We assessed trial quality using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). These criteria included assessment of selection bias, performance bias, attrition bias, detection bias, and reporting bias. We rated each domain separately for each included study in a risk of bias table as showing high risk, unclear risk, or low risk of bias.
We resolved disagreements regarding assessment by discussion and consensus. We did not combine in pooled estimates of effects the results of studies assessed as having high risk of bias and low risk of bias. We noted this under Assessment of risk of bias in included studies.
Measures of treatment effect
In the absence of clinical heterogeneity, we undertook analysis using RevMan 5.3. We used risk ratio (RR) to measure treatment effect for dichotomous outcomes measured in two or more studies evaluating the same intervention. For continuous data, we used mean difference (MD) and the inverse variance method and calculated an overall MD. We used a fixed‐effect model when we found no evidence of significant heterogeneity among studies, and a random‐effects model when heterogeneity was likely (DerSimonian 1986).
Unit of analysis issues
No cluster‐randomized trials met our inclusion criteria. We combined trial data because we had no issues with units of analysis. We planned to include data only from the first phase of cross‐over trials.
Dealing with missing data
All data extracted reflect the original allocation group, enabling an intention‐to‐treat analysis (ITT). We identified drop‐outs when this information was given. We included studies irrespective of whether all or some of the outcome information was available. We presented this information in risk of bias tables (Assessment of risk of bias in included studies) in the selective reporting domain.
In Keller 2000, the data for "oropharyngeal leak pressure" was presented using means and 95% CI, where 95% CI were converted to standard deviation (SD) by the method described in Higgins 2011a. While in Shin 2010, this data was presented as median and 25th and 75th percentiles which were converted to mean and SD by the process described by Wan 2014.
Assessment of heterogeneity
We addressed clinical heterogeneity through the characteristics of included studies (e.g. participant age, weight, height, and BMI; American Society of Anesthesiology (ASA) physical status; duration of surgery). Provided that the data were homogeneous enough to be combined, we performed a meta‐analysis of data from all included studies. We assessed the risk of bias for methodological heterogeneity and summarized this information in Figure 1. To evaluate statistical heterogeneity, we used Chi² and I² tests. We used the Chi² test to evaluate whether heterogeneity exists, while considering P < 0.1 as statistically significant, and we used the I² statistic to quantify inconsistencies. If I² was greater than 75%, we considered heterogeneity to be considerable (Higgins 2011). We used the results of these two tests to assess whether heterogeneity needs to be accounted for in the model. We investigated reasons for any heterogeneity. We used a fixed‐effect model if we found studies to be homogeneous; otherwise we used a random‐effects model.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Assessment of reporting biases
We planned to perform sensitivity analyses for study quality that considered only studies with low or moderate risk of bias, using a risk of bias table for each included study.
We planned to evaluate each study for classification into one of the following categories.
-
Low risk of bias ‐ all criteria met.
-
Moderate risk of bias ‐ one or more criteria unclear.
-
High risk of bias ‐ one or more criteria not applied or met.
We planned to assign low risk of bias when allocation concealment was adequate; moderate risk of bias when allocation concealment was unclear or when study authors did not clearly report the approach; and high risk of bias when allocation concealment was not applied. We planned to evaluate the impact of methodological quality in 'Risk of bias' tables.
Data synthesis
We entered data extracted from the studies into RevMan 5.3 and combined data quantitatively when appropriate. We planned to present primary outcomes as a dichotomous variable (success/no success). We planned to analyse other outcomes as dichotomous or continuous variables. We further planned to calculate standardized mean differences (SMDs) with 95% confidence intervals (CIs). We calculated pooled estimates using the fixed‐effect model unless we observed significant heterogeneity, in which case we used a random‐effects model. For continuous variables, we planned to calculate the mean difference (MD) with 95% CI. We used RR with 95% CI for dichotomous outcomes. We planned to conduct sensitivity analysis to discern the impact of the type of analysis on results.
Subgroup analysis and investigation of heterogeneity
We planned to conduct subgroup analyses to investigate suspected clinical heterogeneity based on:
-
duration of surgery (short vs long);
-
type of surgical procedure (abdominal vs peripheral surgery);
-
positioning during surgery; and
-
differing definitions of the primary outcome.
Sensitivity analysis
We planned to carry out sensitivity analysis according to study methodological quality (for trials with low risk of bias) and missing data (using best‐case and worst‐case scenarios). However, neither was required given that we assessed almost all of the included studies as having high risk of bias and identified no missing data.
'Summary of findings' table and GRADE
We used the principles of the GRADE system (Guyatt 2008) to assess the quality of the body of evidence associated with the following outcomes in our review.
-
Failure to adequately mechanically ventilate according to two variables.
-
Inadequate oxygenation.
-
Inadequate ventilation.
-
-
Time required for effective airway (seconds).
-
Peak airway pressure during positive pressure ventilation.
-
Oropharyngeal leak pressure (seal pressure) (cm H₂O).
We constructed a 'Summary of findings' table (summary of findings Table for the main comparison) using GRADEpro GDT software (GRADEpro GDT 2014).
We graded results according to the GRADE approach (www.bmj.com/advice/sections/shtml) in accordance with recommendations provided in the Cochrane Handbook for Systematic Reviews of Interventions, Chapter 12.2.1 (Higgins 2011).
Results
Description of studies
Results of the search
We identified 4741 records through electronic searches conducted in April 2017. Of these, we excluded 4719 titles and abstracts, as they were duplicates (789) or were irrelevant (3930).
The remaining 22 records qualified for full paper analysis. We identified two additional records by handsearching reference lists of full‐text papers. We therefore accessed 24 full‐text studies and excluded 10 studies after full‐text review, as all of these studies were irrelevant to our review. Of the remaining 14 studies, we identified six RCTs that are awaiting classification.
Five of the RCTs awaiting classification are cross‐over trials, and data related to the first treatment period were not available. We contacted study authors (9 and 10 December 2015, and 1 and 4 April 2017) to request further information. One study is awaiting classification because, although the study was completed in November 2016, the study report has not yet been published. We will assess data from these six studies, when available, and will include studies that meet the review study inclusion criteria in a future update. (See Characteristics of studies awaiting classification; Figure 2.) We included eight studies in this review.
Study flow diagram.
Included studies
We included eight studies with a total of 829 participants (Ambi 2011; Brimacombe 2002; Keller 2000; Khazin 2008; Lu 2002; Natalini 2003a; Natalini 2003b; Shin 2010) (416 participants in the intervention (pLMA) group, and 413 in the control (cLMA) group) of either sex, between 18 and 84 years of age, with American Society of Anesthesiology (ASA) physical status grades I to III (ASA I, II, and III), who were scheduled for elective surgery (i.e. peripheral limb surgery, orthopaedic surgery, minor vascular surgery, peripheral plastic surgery, urological surgery, laparoscopic cholecystectomy, other abdominal and gynaecological laparoscopic procedures). Of these studies Khazin 2008 compared five devices and Shin 2010 compared three airway devices, but we used only data pertaining to cLMA and pLMA.
In one study (Natalini 2003a), participants were obese (BMI > 30). Four studies (Brimacombe 2002; Khazin 2008; Lu 2002; Shin 2010) excluded morbidly obese patients (BMI > 35 kg/m2). Two studies (Ambi 2011; Keller 2000) did not report BMI of participants. Natalini 2003b reported BMI of study participants as 26 ± 6/25 ± 4 pLMA/cLMA.
Five studies (Ambi 2011; Keller 2000; Lu 2002; Natalini 2003a; Shin 2010) used neuromuscular blocking agents for insertion of the intended device and Natalini 2003b used muscle relaxant after LMA placement. Two studies (Brimacombe 2002; Khazin 2008) did not use muscle relaxants in study participants.
Six of the included studies (Ambi 2011; Keller 2000; Khazin 2008; Lu 2002; Natalini 2003a; Natalini 2003b) were single‐centre studies conducted in India, Australia, Israel, and Italy. Shin 2010 was performed at two centres ‐ Japan and South Korea ‐ and Brimacombe 2002 was a multi‐centre trial conducted at eight study sites in seven countries (one each in Australia, Austria, France, Germany, Italy, and Spain; two in the United States).
Among eight included studies, five studies (Ambi 2011; Keller 2000; Natalini 2003a; Natalini 2003b; Shin 2010) did not report funding resources. Brimacombe 2002 reported that some data were sponsored by the Laryngeal Mask Company, which manufactures ProSeal and Classic laryngeal mask airway devices. However, neither the Laryngeal Mask Company nor Dr Brain (the inventor) was involved in the design of the study, data analysis, or manuscript preparation. One of the authors in Lu 2002 had received research funds from Intavent Ltd, the manufacturer of both types of LMAs, but Intavent Ltd did not sponsor this study. Khazin 2008 reported that the study was supported in part by the Joseph Drown Foundation, in Los Angeles, California, in the United States.
Excluded studies
We excluded 10 studies following full‐text assessment owing to inclusion of irrelevant interventions. Of these, five studies (Hosten 2009; Lopez 2011; Seet 2010; Tham 2010; Verghese 2008) compared disposable (LMA ‐ Supreme) versus reusable ProSeal LMA. Francksen 2007 compared three disposable airway devices that were not comparable with cLMA and pLMA. Shariffuddin 2008 compared the Ambu AuraOnce Laryngeal Mask versus the Classic LMA only. Goldmann 2011 compared only pLMA in obese and non‐obese participants. Maltby 2003 compared cLMA/pLMA versus an endotracheal tube, and Brain 2000 compared two types of LMA versus drainage tubes (prototype LMA and new LMA, which were not comparable with cLMA). See Characteristics of excluded studies for details.
Studies awaiting classification
We identified six RCTs that are awaiting classification (Brimacombe 2000; Brimacombe 2001; Cook 2002; Li 2007; Natalini 2003c; NCT02979171). Five of these are cross‐over trials and data related to the first treatment period were not available (Brimacombe 2000; Brimacombe 2001; Cook 2002; Li 2007; Natalini 2003c). We contacted the study authors (9 and 10 December 2015, and 1 and 4 April 2017) to request further information. One study (identifier number: NCT02979171) is awaiting classification. This study was completed in November 2016, but the study report has not yet been published. We will assess these data, when available, and will include studies that meet the review study inclusion criteria in a future update. (See Characteristics of studies awaiting classification.)
Ongoing studies
We identified no ongoing studies.
Risk of bias in included studies
We evaluated the overall quality of studies according to the methods outlined under Assessment of risk of bias in included studies. We present the different bias domains in Characteristics of included studies. We present the graph and summary of risk of bias in included studies in Figure 1 and Figure 3.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Allocation
Investigators described all eight included studies (Ambi 2011; Brimacombe 2002; Keller 2000; Khazin 2008; Lu 2002; Natalini 2003a; Natalini 2003b; Shin 2010) as randomized, and provided sufficient details on the method of randomization.
Regarding allocation concealment, Ambi 2011, Brimacombe 2002, Keller 2000, and Lu 2002 used a sealed envelope technique; Khazin 2008 used computer‐generated codes; and Natalini 2003a and Natalini 2003b used computer‐generated tables. Shin 2010 provided no information about the envelope, and we considered risk to be unclear. Overall, our judgement was that studies were at low risk for selection bias.
Blinding
While assessing performance bias, we believed it was not possible to blind personnel inserting the airway device. Similarly, it was not possible for assessors of the primary outcomes of "inadequate oxygenation and ventilation" to be blinded. However, some studies mentioned that outcome assessors were blinded for some of the secondary outcomes, and other studies described observers as "independent" for some outcomes, which is not the same as being blinded to group allocation.
Brimacombe 2002 stated that an unblinded, trained observer collected data during surgery, and a blinded, trained observer collected data in the post‐anaesthesia care unit (PACU) the following day. This resulted in high risk of detection bias, as investigators collected most of the data during surgery.
We judged Natalini 2003b to be at high risk of bias, as study authors mention that this is an "unblinded" study.
All other studies (Ambi 2011; Keller 2000; Khazin 2008; Lu 2002; Natalini 2003a; Shin 2010) did not provide sufficient information to permit assessment of detection bias. Keller 2000 used unblinded observers for fibre‐optic position and scoring but provided insufficient information regarding other outcomes. Khazin 2008 used independent and blinded observers for only one of the secondary outcomes, and Lu 2002 mentioned that "independent" observers collected the data but did not mention blinding. Ambi 2011, Natalini 2003a, and Shin 2010 did not report anything regarding blinding of personnel and assessors.
On the basis of this information, we judged performance and detection bias to be unclear.
Incomplete outcome data
Six studies (Ambi 2011; Keller 2000; Khazin 2008; Lu 2002; Natalini 2003b; Shin 2010) included in their analyses data on all participants. Of the remaining two studies, Brimacombe 2002 reported loss of 1% intraoperative and 3% postoperative data related to gastric tube placement and position of the drainage tube and airway tube. A total of 53 protocol deviations included in the analysis (as they were minor) were evenly distributed between groups and were related to one of our secondary outcomes ‐ gastric tube insertion. In one of 30 participants in Natalini 2003a, both devices failed; ventilation was managed by face mask, and these participants were excluded from the study. As most of the included studies provided data on all study participants, and the other two studies (Brimacombe 2002; Natalini 2003a) had very low rates of attrition, we assigned low risk of bias for this domain.
Selective reporting
Six of eight studies (Brimacombe 2002; Keller 2000; Lu 2002; Natalini 2003a; Natalini 2003b; Shin 2010) reported data for all outcomes assessed. Ambi 2011 did not report postoperative nausea and vomiting, but this is not one of our major outcomes of interest. Khazin 2008 mentioned that researchers would monitor plateau pressure, inspired and expired minute volume, SaO₂, ETCO₂, adequacy of ventilation, airway manipulations, number of failed insertion attempts, gastric insufflations due to positive pressure ventilation, and complications. In the results section, study authors stated that "respiratory variables were similar in both groups," which covers all of the variables mentioned above. Owing to this, we have assigned low risk of reporting bias.
Other potential sources of bias
We found no other potential sources of bias in the included studies.
Effects of interventions
We have detailed important outcomes in summary of findings Table for the main comparison. In relation to effective airway time (Analysis 3.1), it is important to note that the standard deviation (SD) of most included studies suggests a right‐skewed distribution of sample values, suggesting imprecision in the results.
Outcomes
Primary outcomes
Failure to mechanically ventilate adequately
Inadequate oxygenation
Seven trials reported oxygenation as assessed by peripheral arterial oxygen saturation (SaO₂) (Ambi 2011; Brimacombe 2002; Khazin 2008; Lu 2002; Natalini 2003a; Natalini 2003b; Shin 2010). Two studies reported "hypoxic events" rather than a measurement point below a stated threshold (Ambi 2011; Shin 2010); and Brimacombe 2002 reported SaO₂ as "less than 90%." Lu 2002 considered oxygenation as "suboptimal" if SaO₂ fell to 94% to 90%, and as "failed oxygenation" if less than 90%. Two trials (Natalini 2003a; Natalini 2003b) did not report quantitative results ‐ only that "SaO₂ was similar in both groups." Khazin 2008 reported that "respiratory variables were similar in both groups."
This analysis included data from four studies (N = 617 participants; 310 allocated to pLMA and 307 to cLMA) (Ambi 2011; Brimacombe 2002; Lu 2002; Shin 2010). Of these, three studies (N = 233; 118 participants in the pLMA group and 115 in the cLMA group) reported no hypoxic events in either group (Ambi 2011; Lu 2002; Shin 2010). Only Brimacombe 2002 reported actual hypoxia events (SaO₂ < 90%) (N = 384; pLMA and cLMA groups = 192/192). Data show seven hypoxic events (1.8%) with no clinically important differences between groups. Three of 192 participants (1.6%) in the pLMA group and four of 192 (2.1%) in the cLMA group experienced hypoxic events. Owing to limited available data, information was insufficient to show the effects of pLMA and cLMA on hypoxic events (RR 0.75, 95% CI 0.17 to 3.31; low‐quality evidence; Analysis 1.1).
It was uncertain whether four studies that looked at inadequate oxygenation showed any difference between pLMA and cLMA groups (Ambi 2011; Brimacombe 2002; Lu 2002; Shin 2010). Furthermore, results from these four studies are imprecise; only Brimacombe 2002 reported positive events. We assessed three of these four studies as adequately concealing allocation to study arms, and we assessed Shin 2010 as providing unclear information. We noted no selective reporting in any of these studies. Blinding of participants was not clear in all four studies, and investigators blinded assessors to outcomes (Brimacombe 2002) or provided unclear information (Ambi 2011; Lu 2002; Shin 2010). We therefore downgraded the quality of evidence by three levels from high to very low.
Inadequate ventilation
Only Lu 2002 reported the incidence of inadequate ventilation (N = 80; 40/40 pLMA/cLMA groups). Lu 2002 defined inadequate ventilation as end‐tidal carbon dioxide (ETCO₂) greater than 45 mm Hg to 55 mm Hg (> 6.0 to 7.3 kPa), and failed ventilation as ETCO₂ greater than 55 mm Hg (> 7.3 kPa). No participants experienced any episodes of inadequate ventilation.
Secondary outcomes
Inability to insert LMA or abandonment of intended device insertion
Seven studies (N =769; 386 in pLMA group and 383 in cLMA group) looked at successful or failed laryngeal mask airway (LMA) placement (Ambi 2011; Brimacombe 2002; Keller 2000; Khazin 2008; Lu 2002; Natalini 2003a; Shin 2010). Researchers assessed proper placement by looking at capnograph wave form (Brimacombe 2002; Khazin 2008; Lu 2002; Shin 2010); chest movement (Khazin 2008; Lu 2002; Shin 2010); ETCO₂ (Shin 2010); no audible leak (Brimacombe 2002); or leak fraction greater than 15% (Natalini 2003a); or by checking position by using a fibre‐optic scope (Keller 2000). Two studies did not define an effective airway (Ambi 2011; Khazin 2008). Four studies used two or more techniques or modalities to monitor proper placement (Brimacombe 2002; Khazin 2008; Lu 2002; Shin 2010). After three attempts, all studies reported using an alternate device.
Five studies (Ambi 2011; Keller 2000; Khazin 2008; Lu 2002; Shin 2010; N = 325) reported no events in either arm. Two studies (Brimacombe 2002; Natalini 2003a; N = 444; 222/222 pLMA/cLMA groups) reported failure events. Of these, seven participants experienced failure of insertion events (1.6%): five from pLMA groups (2.25%) and two from cLMA groups (0.9%). Meta‐analysis of all seven studies showed no important differences between groups (RR 2.20, 95% CI 0.50 to 9.72; I² = 48%; P = 0.30; Analysis 2.1).
Time to establish an effective airway (seconds)
Two studies observed time to achieve an effective airway, as measured by taking the device in hand until the effective airway was achieved (Ambi 2011; Brimacombe 2002; N = 434; 217/217 pLMA/cLMA groups). Meta‐analysis showed that time required for insertion was longer with pLMA than with cLMA (MD 10.12, 95% CI 5.04 to 15.21; I² = 0%; P < 0.0001; Analysis 3.1). We downgraded evidence quality by two levels from high to low because researchers did not blind participants and outcome assessors (Brimacombe 2002), or we assessed risk as unclear (Ambi 2011). Data analysis showed imprecise results owing to high standard deviations (SDs) in three of four instances.
Failure to insert gastric tube
Five studies investigated gastric tube insertion (Ambi 2011; Brimacombe 2002; Lu 2002; Natalini 2003a; Natalini 2003b). Three studies compared pLMA and cLMA groups (Brimacombe 2002; Natalini 2003a; Natalini 2003b). Two studies (Ambi 2011; Lu 2002) considered gastric tube insertion in the pLMA group. Investigators in Natalini 2003a and Natalini 2003b inserted a gastric tube before inserting the airway device in the cLMA group. Brimacombe 2002 attempted orogastric tube insertion in 168 participants (95/73 in pLMA/cLMA groups). The failure rate was much lower in the pLMA group (11 participants (11.6%) vs 33 participants (45%)). Insertion time was also shorter in the pLMA group (22 ± 18 seconds vs 38 ± 56 seconds).
Peak airway pressure (cm H₂O)
Four studies assessed peak airway pressure during positive pressure ventilation (Khazin 2008; Lu 2002; Natalini 2003a; Natalini 2003b; N = 259; 130/129 pLMA/cLMA groups). Meta‐analysis showed lower pressure in the cLMA group (MD 0.84, 95% CI 0.02 to 1.67; I² = 0%; P = 0.04; Analysis 4.1).
Blinding of participants and outcome assessors was not clear in three studies (Khazin 2008; Lu 2002; Natalini 2003a), and Natalini 2003b was unblinded for both participants and assessors. We therefore downgraded the quality of evidence from high to low.
Oropharyngeal leak pressure (seal pressure) (cm H₂O)
Six studies assessed oropharyngeal leak pressure (Ambi 2011; Brimacombe 2002; Keller 2000; Khazin 2008; Lu 2002; Shin 2010; N = 709; 356/353 pLMA/cLMA groups). Five studies (Ambi 2011; Brimacombe 2002; Keller 2000; Lu 2002; Shin 2010) determined seal pressure by using gas flow of three to five litres per minute and closing the adjustable pressure limiting (APL) valve at 40 cm H₂O; Khazin 2008 checked this at 60 cm H₂O. Mean oropharyngeal leak pressure was higher in the pLMA group (MD 6.93, 95% CI 4.23 to 9.62; I² = 87%; P < 0.00001; Analysis 5.1.
Blinding of participants was unclear in all studies; Brimacombe 2002 did not blind outcome assessors, and remaining five studies (Ambi 2011; Keller 2000; Khazin 2008; Lu 2002; Shin 2010) provided unclear information. Furthermore, data analysis revealed high levels of heterogeneity (inconsistency); we therefore downgraded the quality of evidence from high to low.
Complications
Mucosal injury (blood staining)
Four studies reported mucosal injury (Ambi 2011; Brimacombe 2002; Lu 2002; Shin 2010; N = 617; 310/307 pLMA/cLMA). A total of 45 participants (14.5%) in the pLMA group and 30 (9.7%) in the cLMA group had mucosal injury. Although the incidence is 4.8% higher in the pLMA group, we are uncertain that this difference is of any significance (RR 1.49, 95% CI 0.97 to 2.28; I² = 0%; P = 0.07; Analysis 6.1).
Sore throat
Five studies reported this outcome (Ambi 2011; Brimacombe 2002; Natalini 2003a; Natalini 2003b; Shin 2010; N = 653; 329/324 pLMA/cLMA groups). Overall, sore throat occurred in 49 participants (14.9%) in the pLMA group and in 63 (19.4%) in the cLMA group. The risk ratio of sore throat was 4.5% higher in the cLMA group but there is uncertainty as to whether this difference is of any significance (RR 0.77, 95% CI 0.55 to 1.08; I² = 22%; P = 0.12; Analysis 6.2).
Bronchospasm or laryngospasm
Three studies looked at bronchospasm (Ambi 2011; Brimacombe 2002; Shin 2010; N = 537; 270/267 pLMA/cLMA groups). Brimacombe 2002 reported bronchospasm in one of 192 participants (0.5%) in the pLMA group. Meta‐analysis showed no important differences between pLMA and cLMA groups (RR 3.00, 95% CI 0.12 to 73.19; I² = 0%; P = 0.50; Analysis 6.3).
No studies reported any episodes of laryngospasm.
Gastric insufflation
Two studies looked at gastric insufflation (Brimacombe 2002; Shin 2010; N = 487; 245/242 pLMA/cLMA groups). Brimacombe 2002 reported gastric insufflation in one of 192 participants (0.5%) in the pLMA group and in three of 192 participants (1.5%) in the cLMA group. Here also, we are uncertain whether this small difference of 1% between the devices is of any significance (RR 0.33, 95% CI 0.03 to 3.18; I² = 0%; P = 0.34; Analysis 6.4).
Regurgitation or aspiration
Four studies reported regurgitation events (Ambi 2011; Brimacombe 2002; Khazin 2008; Shin 2010; N = 597 participants; 300/297 pLMA/cLMA groups). Khazin 2008 reported a total of nine events of regurgitation based on hypopharyngeal pH < 4, in four of 30 pLMA group participants (13.3%) and in five of 30 cLMA participants (16.7%). Meta‐analysis showed no important differences between pLMA and cLMA groups (RR 0.80, 95% CI 0.24 to 2.69; I² = 0%; P = 0.72; Analysis 6.5). No studies reported any episodes of aspiration.
Coughing
Two studies investigated the occurrence of coughing (Brimacombe 2002; Khazin 2008; N = 444; 222/222 pLMA/cLMA groups). Khazin 2008 reported no coughing events, and Brimacombe 2002 reported coughing events in 13 of 192 participants (6.8%) in the pLMA group and in 18 of 192 participants (9.4%) in the cLMA group. Data show a slightly increased risk (2.6%) of cough among cLMA group participants. Meta‐analysis showed no important differences between groups (RR 0.72, 95% CI 0.36 to 1.43; I² = 0%; P = 0.35; Analysis 6.6).
Excessive leak
Two studies looked at excessive leak (Natalini 2003a; Natalini 2003b; N = 120; 60/60 pLMA/cLMA groups). Meta‐analysis showed no important differences between devices (RR 3.00, 95% CI 0.13 to 70.83; I² = 0%; P = 0.5; Analysis 6.7).
Obstruction related to two devices
Brimacombe 2002 reported one (0.5%) and three (1.5%) events of airway obstruction among 192 participants in the pLMA and cLMA groups, respectively (Brimacombe 2002).
Postoperative nausea and vomiting
No included studies reported postoperative nausea and vomiting as an outcome.
Discussion
Summary of main results
The eight included studies enrolled a total of 829 participants.
Of four studies that reported adequate oxygenation, three reported no hypoxia in either group (Ambi 2011; Lu 2002; Shin 2010). The fourth study (Brimacombe 2002) reported actual events of hypoxia. Owing to limited available data, information was insufficient to show the effect on hypoxic events (risk ratio (RR) 0.75, 95% confidence interval (CI) 0.17 to 3.31; low‐quality evidence).
Lu 2002 looked at inadequacy of ventilation and reported no episodes of inadequate ventilation in either group.
Results from meta‐analyses show statistically longer time to achieve an effective airway in ProSeal laryngeal mask airway (pLMA) group participants (Ambi 2011; Brimacombe 2002) and higher mean oropharyngeal leak pressure (Ambi 2011; Brimacombe 2002; Keller 2000; Khazin 2008; Lu 2002; Shin 2010). Differences in these outcomes, although statistically important, have little impact on clinical practice. Analysis of four studies that compared peak airway pressure (Khazin 2008; Lu 2002; Natalini 2003a; Natalini 2003b) revealed higher pressures in the pLMA group. Data from seven studies looking at failure to insert the intended device or use of an alternate device (Ambi 2011; Brimacombe 2002; Keller 2000; Khazin 2008; Lu 2002; Natalini 2003a; Shin 2010) show no significant differences between pLMA and Classic laryngeal mask airway (cLMA) device groups.
Results were uncertain in terms of differences for several complications, including mucosal injury (Ambi 2011; Brimacombe 2002; Lu 2002; Shin 2010), sore throat (Ambi 2011; Brimacombe 2002; Natalini 2003a; Natalini 2003b; Shin 2010), bronchospasm (Ambi 2011; Brimacombe 2002; Shin 2010), gastric insufflation (Brimacombe 2002; Shin 2010), regurgitation (Ambi 2011; Brimacombe 2002; Khazin 2008; Shin 2010), coughing (Brimacombe 2002; Khazin 2008), and excessive leak (Natalini 2003a; Natalini 2003b). Analysis to assess obstruction related to use of the devices was not possible because only one study looked at this outcome (Brimacombe 2002). No studies reported postoperative nausea and vomiting as an outcome.
Overall completeness and applicability of evidence
The numbers of included studies and participants were modest (eight studies, 829 participants), and most studies were unblinded or unclear to blinding for most of the important outcomes. Given the nature of both the intervention and the main outcomes, this is probably unavoidable in practice. Most included studies reported no events, so even fewer trials were eligible for analysis.
Half of the included studies assessed the primary outcomes of this review (four assessed adequacy of oxygenation, and one, adequacy of ventilation). However, most of these studies reported no events in pLMA or cLMA groups (three for oxygenation, and one for ventilation); one study reported adequacy of oxygenation events (Brimacombe 2002). No studies reported adequacy of ventilation events.
The evidence was not sufficiently robust to enable reasonable conclusions. In current anaesthesia practice, use of endotracheal intubation for airway maintenance is being rapidly replaced by use of supraglottic devices. Assessment of the adequacy of mechanical ventilation through adequate oxygenation and ventilation is more relevant; however, these are comparable for both of the devices assessed in this review.
Quality of the evidence
We included eight somewhat clinically heterogenous trials that involved a total of 829 participants. Low numbers of participants in both the intervention (pLMA) group (N = 416) and the control (cLMA) group (N = 413) rendered reliable comparisons difficult for most of our important clinical outcomes.
The overall quality of evidence was low. Seven of eight studies used reliable methods of randomization and concealment. One study (Natalini 2003b) did not blind participants, personnel, or outcome assessors, and two studies reported blinding both participants and outcome assessors but only for some secondary outcomes (Brimacombe 2002; Khazin 2008).
Most trials reported no events in either arm. This may have occurred because the events are not common, but the data could not be used to conclude in favour of either device. Furthermore, heterogeneity and imprecision contributed to the overall low quality of evidence. These issues warrant additional research to arrive at meaningful conclusions.
Potential biases in the review process
We made changes to outcomes presented in the protocol for this review after extensive discussion among review authors, peer referees, and editors. We made these changes with the aim of making the review more meaningful in terms of clinical practice (see Differences between protocol and review). However, modifications to primary and secondary outcomes may have introduced bias into the review because we applied these changes after the initial search and after review of outcomes reported by the included studies.
Most studies were unclear regarding blinding of both those performing the intervention and those assessing responses. This variation in study design may also have introduced bias.
Another important factor that could have been responsible for bias was that different studies measured outcomes differently ‐ sometimes with large variations ranging from subjective clinical assessment to objective instrumental assessment.
Outcomes reported in the 'Summary of findings' table have been modified, as described above, (also see Differences between protocol and review). We also included "time required for an effective airway" as an outcome based on the review of included studies. Both these could have been a source of bias.
Agreements and disagreements with other studies or reviews
We found three other systematic reviews that addressed the research question of this review (Brimacombe 2002a; Cook 2005; Zhang 2012). We cannot compare Zhang 2012 with our review because Zhang considered children only; Brimacombe 2002a and Cook 2005 included adults.
Findings of two reviews (Brimacombe 2002a; Cook 2005) were consistent with the findings of our review in that the time required for insertion of pLMA was longer than for cLMA. Overall success rates were similar with both devices; seal pressure was better with pLMA.
Brimacombe 2002a and Cook 2005 have reported on various complications related to the use of these devices. Neither of these reviews have performed meta‐analyses, which makes it inappropriate to compare the devices. However, they have reported some comparisons based upon composite and individual study data. The incidence of mucosal injury/airway trauma is higher in the pLMA group when compared to cLMA. Regurgitation without aspiration is higher in cLMA when higher pressures are used during positive pressure ventilation. The incidence of gastric insufflation was similar with both the devices when the drainage tube was not used, otherwise the incidence was lower in the pLMA group.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Study flow diagram.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Comparison 1 Failure to mechanically ventilate, Outcome 1 Inadequate oxygenation.
Comparison 2 Failure to insert or use of alternate device, Outcome 1 Failure to insert or use of alternate device.
Comparison 3 Effective airway time (seconds), Outcome 1 Effective airway time (seconds).
Comparison 4 Peak airway pressure during positive pressure ventilation, Outcome 1 Peak airway pressure during positive pressure ventilation.
Comparison 5 Oropharyngeal leak pressure (seal pressure), Outcome 1 Oropharyngeal leak pressure (seal pressure) (cm H₂O).
Comparison 6 Complications, Outcome 1 Mucosal injury.
Comparison 6 Complications, Outcome 2 Sore throat.
Comparison 6 Complications, Outcome 3 Bronchospasm.
Comparison 6 Complications, Outcome 4 Gastric insufflation.
Comparison 6 Complications, Outcome 5 Regurgitation.
Comparison 6 Complications, Outcome 6 Coughing.
Comparison 6 Complications, Outcome 7 Excessive leak.
| ProSeal versus Classic laryngeal mask airway (LMA) for positive pressure ventilation in adult patients undergoing elective surgery | ||||||
| Patient or population: adult patients (18 years of age and older) for positive pressure ventilation undergoing elective surgery Comparison: Classic laryngeal mask airway (cLMA) | ||||||
| Outcomes | Illustrative comparative risks* (95% CI) | Relative effect | Number of participants (studies) | Quality of the evidence | Comments | |
| Assumed risk | Corresponding risk | |||||
| cLMA | pLMA | |||||
| Inadequate oxygenation (SaO₂ < 90% on peripheral oxygen saturation) | 13 per 1000 | 10 per 1000 (2 to 43) | RR 0.75 (0.17 to 3.31) | 617 (4 studies) | ⊕⊝⊝⊝ | |
| Inadequate ventilation ( ETCO₂ > 45 mm Hg) | See comment | See comment | Not estimable | 80 | See comment | No episodes of inadequate ventilation in either group |
| Time required for effective airway (seconds) | Mean effective airway time in the control group ranged from 22.24 seconds to 31 seconds | Mean effective airway time in the intervention group was 10.12 seconds longer (95% CI 5.04 to 15.21) | 434 | ⊕⊕⊝⊝ Lowb | Outcome added post protocol as some included studies looked at this outcome. | |
| Peak airway pressure during positive pressure ventilation (cm H₂O) | Peak airway pressure in the control group ranged from 17 cm H₂O to 24.5 cm H₂O | Mean peak airway pressure during positive pressure ventilation was higher in the pLMA group by 0.84 cm H₂O (95% CI 0.02 to 1.67) | 259 (4) | ⊕⊕⊝⊝ Lowc | ||
| Oropharyngeal leak pressure (seal pressure) (cm H₂O) | Oropharyngeal leak pressure in the control group ranged from 14.64 cm H₂O to 26 cm H₂O | Mean oropharyngeal leak pressure (seal pressure) was higher in the pLMA group by 7.17 cm H₂O (95% CI 4.85 to 9.5) | 709 | ⊕⊕⊝⊝ | ||
| Sore throat | 194 per 1000 | 150 per 1000 (107 to 210) | RR 0.77 (0.55 to 1.08) | 653 (5) | ⊕⊕⊝⊝ Lowe | |
| Regurgitation | 17 per 1000 | 13 per 1000 (4 to 45) | RR 0.8 (0.24 to 2.69) | 597 (4) | ⊕⊕⊝⊝ Lowf | |
| *The basis for the assumed risk (e.g. median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence. | ||||||
| aThree studies reporting this outcome were assessed as demonstrating adequate allocation concealment and one was unclear. Blinding of participants, personnel and outcome assessor were assessed as unclear in three studies and one study was unblinded. Furthermore, the results are imprecise, as only one out of four studies has reported positive events, therefore evidence is downgraded three levels. bData from two studies with imprecise results resulted in downgrading of evidence by two levels. cBlinding of participants, personnel and outcome assessors were unclear in three studies and one study was unblinded. This resulted in downgrading the evidence by two levels. dBlinding of participant and personnel were unclear in all six studies while blinding of outcome assessor was not conducted in one study and was unclear in the remaining five studies. Data analysis indicated high levels of heterogeneity (inconsistency). For these reasons, the evidence was downgraded two levels. eBlinding of participant and personnel were unclear in four studies and was not done in one study, while blinding of outcome assessor was not conducted in two studies and was unclear in the remaining three studies. This resulted in downgrading the evidence by two levels. fBlinding of participant and personnel were unclear in all four studies while blinding of outcome assessor was not conducted in one study and was unclear in the remaining three studies, therefore evidence was downgraded two levels. | ||||||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Inadequate oxygenation Show forest plot | 4 | 617 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.75 [0.17, 3.31] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Failure to insert or use of alternate device Show forest plot | 7 | 769 | Risk Ratio (M‐H, Fixed, 95% CI) | 2.2 [0.50, 9.72] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Effective airway time (seconds) Show forest plot | 2 | 434 | Mean Difference (IV, Fixed, 95% CI) | 10.12 [5.04, 15.21] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Peak airway pressure during positive pressure ventilation Show forest plot | 4 | 259 | Mean Difference (IV, Fixed, 95% CI) | 0.84 [0.02, 1.67] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Oropharyngeal leak pressure (seal pressure) (cm H₂O) Show forest plot | 6 | 709 | Mean Difference (IV, Random, 95% CI) | 6.93 [4.23, 9.62] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Mucosal injury Show forest plot | 4 | 617 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.49 [0.97, 2.28] |
| 2 Sore throat Show forest plot | 5 | 653 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.77 [0.55, 1.08] |
| 3 Bronchospasm Show forest plot | 3 | 537 | Risk Ratio (M‐H, Fixed, 95% CI) | 3.0 [0.12, 73.19] |
| 4 Gastric insufflation Show forest plot | 2 | 487 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.33 [0.03, 3.18] |
| 5 Regurgitation Show forest plot | 4 | 597 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.8 [0.24, 2.69] |
| 6 Coughing Show forest plot | 2 | 444 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.72 [0.36, 1.43] |
| 7 Excessive leak Show forest plot | 2 | 120 | Risk Ratio (M‐H, Fixed, 95% CI) | 3.0 [0.13, 70.83] |
