A Randomized Controlled Trial Comparing the Cuffed Oropharyngeal Airway and the Laryngeal Mask Airway in Spontaneously Breathing Anesthetized Adults 

Four hundred fifty-three consenting adult patients undergoing general anesthesia for routine procedures at three institutions were randomly assigned using a 2:1 allocation to have either the COPA (n = 302) or LMA (n = 151) used for airway management (Table 1). The study sites-Johns Hopkins Medical Institution (JHMI; Baltimore, Maryland), Cairns Base Hospital (CBH; Cairns, Australia), and Nambour General Hospital (NGH; Nambour, Australia)-were chosen to provide anesthetists of varying seniority and with a broad spectrum of clinical experience and expectation of the investigational devices. Clinical investigation committee approval was obtained from all three study sites. Patients were excluded from the trial if they were younger than 18 yr, were pregnant, had a history of gastroesophageal reflux, a known or predicted difficult airway (e.g., Mallampati classification 4), or were considered unsuitable for a face mask or LMA. All participating anesthetists viewed an instructional video produced by the research team that informed them of the study protocol and prescribed technique for use and placement of both devices. Patients were assessed by trained data collectors (different from the anesthetist) regarding tolerance of the devices under anesthesia (heart rate [based on electrocardiogram], blood pressure [sphygmomanometer or automated blood pressure manometer], oxygen saturation [pulse oximetry], tidal volume [in-line spirometer of anesthesia machine], end-tidal carbon dioxide [circuit end-tidal monitor]), use of airway supportive measures, and clinical events associated with each device.

The initial size of the COPA was chosen by placing the distal tip of the device at the angle of the jaw of a patient in the supine position with the COPA straight up, perpendicular to the floor. When viewed from the side, the tooth/lip guard of the COPA would be about 1 cm ventral to the lip in a device of the proper size. The initial size of the LMA was chosen according to the manufacturer's weight-related recommendations. [2] 

Anesthetic management was standardized according to the following protocol: Monitoring was applied before induction and included an electrocardiograph, pulse oximeter, capnograph, noninvasive blood pressure monitor, and tidal volume monitor. Sedation (with 0–2 mg midazolam and 0–100 micro gram fentanyl given intravenously) was administered at the discretion of the anesthetist. After at least 30 s of preoxygenation, anesthesia was induced using 1.5–2 mg/kg propofol given intravenously. Additional 40-mg boluses of propofol could be given at the discretion of the anesthetist until an adequate level of anesthesia was obtained. After loss of the lid reflex, the LMA or the COPA was inserted using the prescribed technique. The cuff was inflated with air as necessary to establish an effective airway. Each insertion attempt (defined as a forward and backward movement of the device) was recorded. At the second and third insertion attempts, the anesthetist was free to try an alternate size. If the anesthetist could not effectively establish an airway using the initial randomized device after three insertion attempts, then the alternative device was used. If the anesthetist could not establish an effective airway using the second device within three attempts at insertion, the patient was managed as clinically indicated.

Anesthesia was maintained with a propofol infusion (50–200 micro gram [center dot] kg sup -1 [center dot] min sup -1) and nitrous oxide (up to 70%) as clinically indicated. Cuff volumes, resultant pressures (blinded to the anesthetist), and airway leak pressures were measured when the anesthetist obtained an acceptable airway immediately after placement of the device and recorded. Leak pressure was determined by closing the expiratory valve of the circuit and noting the airway pressure (measured in the anesthesia circuit) at which oropharyngeal gas leak occurred into the mouth.

Airway manipulations (head tilt, chin lift, jaw thrust, continuous positive airway pressure) were used as necessary and recorded. The time required to achieve a clinically adequate airway not requiring manual support and the time of resumption of spontaneous ventilation was recorded. If problems occurred, they were noted, timed, and an explanation given. Adverse events were classified into major (aspiration, regurgitation, laryngospasm, wheezing, failure of device after three attempts, use of succinylcholine, use of any alternative device for airway rescue, oxygen saturation < 92%), minor intraoperative (coughing, gagging, hiccup, patient movement, stridor, blood detected on device, need for continuous airway support, nasogastric tube insertion to decompress air in the stomach), and immediate and next-day postoperative (shivering, sore throat, lip trauma, hearing changes, ear pain, jaw-neck-mouth ache, postoperative vomiting). Vital signs were recorded before induction, at 5-min intervals for the first 15 min or until surgical incision, 5 min after surgical incision, and every 15 min until the device was removed, and once after removal.

Subjective assessment of various aspects of using the devices were solicited from the anesthetists using a scale that included the parameters excellent, good, fair, and poor. This included the quality of the patient's airway before the commencement of surgery, ease of device insertion, ease of airway management, effectiveness of sizing technique, ease of hands-free anesthesia delivery, and overall usefulness of the device used. Before discharge and on the following day, the patient was questioned for the presence of sore throat, hearing changes, vomiting, ear pain, or lip swelling. Any additional comments were also recorded.

The primary statistical objective of the trial was to determine if the adverse event rates in the two treatments (COPA or LMA) were equivalent. The definition of equivalence was +/- 10% with a background rate of 15%. The sample size was chosen to yield 90% power to detect nonequivalent event rates. Randomization, stratified by institution, using permuted blocks (blinded to anesthetists) was used to assign individual patients to treatment groups in a ratio of 2:1 (COPA:LMA). The randomization was unbalanced to yield more clinical exposure with the COPA because this was the newer device. Group comparisons and all other analyses were made based on randomized treatment assignment (intention to treat). The association between categorical variables was tested using the chi-squared test. The magnitude of differences in categorical variable groups between groups was estimated using the odds ratio and 95% confidence limits. Differences between the treatment groups with respect to continuously varying measures were tested using a random effects linear model. For non-normally distributed measures, comparisons were based on logarithmically transformed values or nonparametric tests were performed. Unless otherwise noted, data are presented as means +/- SD. Probability values < 0.05 were considered significant.

(Table 1) shows demographic, surgical, and anesthetist details. Randomization was successful in balancing the treatment groups with respect to study site, sex, history of smoking, height, weight, age, Mallampati score, American Society of Anesthesiologists physical status, type or duration of surgical procedure, training level of the anesthetist placing the device, or induction or maintenance anesthetic management. There were no significant differences between groups with respect to the randomized airway device and the final airway device used (Table 2). Cardiorespiratory data (Table 3) revealed no significant differences between the two groups.

Sixty-two anesthetists participated in the trial, reflecting a wide range of clinical experience (range, 1–30 yr). Protocol deviations were most common at JHMI (19 of 20) and occurred more frequently (P < 0.02) in patients assigned to the COPA group (18 of 302) than in those in the LMA group (2 of 151). Protocol deviations were included in the analysis.

Insertion was evaluated as easy for both devices. First-time successful insertion rates, taken alone, were 81% for the COPA and 89% for the LMA (P = 0.05). Success on the second and third attempts was similar, as was the percentage of insertion attempts for each device as the second, crossover device (Table 2). Groups were not different when comparing all insertion attempts (P = 0.26). Failure of a randomized device to be effective for airway management was similar in both groups (COPA, 5.4%; LMA, 3.4%; P = 0.31).

The COPA group had a higher cuff volume (COPA, 34 compared with LMA, 25 ml; P = 0.0001) but a lower cuff pressure (COPA, 93 cm H sub 2 O compared with LMA, 129 cm H²O; P = 0.0001) and leak pressure (COPA, 17 cm H²O compared with LMA, 20 cm H²O; P = 0.0003). There were significantly more airway manipulations with the COPA: head tilt (COPA, 54.2% compared with LMA, 3.3%; P < 0.001), chin lift (COPA, 42.8% compared with LMA, 1.3%; P < 0.001) and jaw thrust (COPA, 4.7% compared with LMA, 0.7%; P = 0.025). One or two airway manipulations were required in 35% of patients with a COPA compared with only 4% for the LMA, and 36% of patients with the COPA required more than two brief airway manipulations compared with none for patients in whom the LMA was used (P = 0.001;Table 4). Similarly, continuous chin support was used in nearly 30% of patients with the COPA, whereas this was used in no patients with the LMA (P < 0.0005;Table 4). Although airway quality after device placement was most commonly considered excellent with the LMA (76.7%), there was a broader range of performance when using the COPA (fair, 22.6%; good, 45.1%; excellent, 25.6%).

The median time to achieve an effective airway was longer with the COPA than with the LMA (COPA, 150 s compared with LMA, 106 s; P = 0.004). Manual support was required to establish an effective airway in 29 of 302 (9.6%) patients with the COPA, in contrast to only 1 of 150 (0.66%) patients with the LMA (P < 0.001). Time to spontaneous breathing was similar for the COPA and LMA (COPA, 7.3 compared with LMA, 6.7 min; P = 0.44). During the apneic period it was possible to ventilate all patients until spontaneous breathing resumed. Although no patient suffered any adverse clinical consequence from participating in this study, at least one potentially serious event occurred in 43 of 302 (14.2%) patients in the COPA group and in 18 of 151 (11.9%) patients in the LMA group; this was not different between the groups (P = 0.50;Table 5). This includes one patient, managed with a COPA, who, during repositioning from the supine to the lateral position, coughed and probably aspirated bile-stained fluid. Regurgitation occurred in one additional patient in the COPA group and in another in the LMA group. Each case was probably related to inadequate depth of anesthesia. Forty-four patients (COPA, 30; LMA, 14; P = 0.823) were noted to have hemoglobin desaturation (SaO²< 92%).

The overall number of patients with minor intraoperative problems was similar (COPA, 24.2% compared with LMA, 30.5%; P = 0.15;Table 5(a)), but hiccup occurred more frequently with the LMA (COPA, 1.7% compared with LMA, 5.3%; P = 0.03;Table 6). When the devices were removed, blood was detected on the LMA more frequently than on the COPA (LMA, 15.3% compared with COPA, 5.8%; P < 0.001; odds ratio [95% confidence limits], 3.01 [1.56–5.83]). Overall, the number of patients with immediate (P < 0.0001) and next-day (P = 0.02) postoperative problems was greater in the LMA group (Table 5). Specifically, sore throat was significantly more frequent with the LMA than with the COPA, both in the immediate postoperative period (COPA, 4.7% compared with LMA, 21.9%; P < 0.0001; odds ratio, 5.75 [2.97–11.1]) and at the next day followup (COPA, 8.4% compared with LMA, 16.1%; P < 0.01; odds ratio, 2.09 [1.15–3.81]). Although reports of jaw, neck, and mouth ache were not significant between groups (COPA, 10 compared with LMA, 1; P = 0.08), they were associated with > 2 or continuous airway manipulation (P = 0.0002). When postoperative ache (jaw, neck, chin), associated with continuous chin lift in the COPA group, is included with all other after effects, late postoperative pain and discomfort becomes similar in both groups (COPA, 11.2% compared with LMA, 15.9%; P = 0.24). Overall, the total number of patients with any problem (major intraoperative, minor intraoperative, immediate or next-day postoperative) was higher in the LMA group than the COPA group (LMA, 56.3% compared with COPA, 40.7%; P <0.002;Table 5). Table 7shows a summary of problems by study site.

Ease of insertion was evaluated by the anesthetists as similar between the two devices (P = 0.26), but airway quality (P <0.001), sizing technique (P < 0.001), ease of attachment of the airway device (P < 0.001), and overall usefulness (P < 0.001) were determined to be better with the LMA than with the COPA.

This study demonstrates some similarities between the COPA and LMA that are relevant to anesthetic practice. In general, both devices were effective in establishing an airway for spontaneously breathing adults under general anesthesia. Both devices were easy to insert, tolerated well by patients under general anesthesia, and not associated with a high frequency or level of major clinical problems. On this basis the devices are equivalent, but this study also revealed some important differences between the devices that are clinically significant for both the patient and anesthetist.

The frequency of events that could lead to significant complications was higher in this study than in previous trials investigating both the LMA [3–5]and the COPA. [1]The incidence of hypoxia (SaO²< 90%) with the LMA is generally less than 1.5%, [3,5,6]compared with 7.2%(SaO²< 92%) in the current study. In a previous study of 100 patients, the incidence of hypoxia (SaO²<90%) with the COPA was 6% compared with 9.6%(SaO²< 92%) for the current study. The factors producing a higher event rate may include different thresholds (< 90% vs. < 92%) for identifying clinical events, use of intravenous rather than volatile anesthesia, interanesthetist variability in experience with each device, and strict adherence to an intention-to-treat analysis (e.g., inclusion of hypoxic events before placement of any device). In addition, calculation and comparison of event rates with the number of patients with problems between the groups must be made cautiously because several events (e.g., laryngospasm, hypoxemia, use of succinylcholine, and failed use) may, in fact, be linked.

Only 2 of the 62 participating anesthetists had ever used the COPA before. Although all of the anesthetists had previously been exposed to the LMA, their clinical experience with the device varied, but was higher in the Australian sites than the U.S. site. Lopez-Gil et al. [5]showed that there is a short-term learning curve with LMA use in at least pediatric anesthetic practice and that the average problem rate per patient could be reduced from 62% to 2% within 75 uses. It is not yet clear whether such a learning curve exists for the COPA, but a previous study investigating the first 50 uses by each of two experienced anesthetists showed evidence of differences in performance between the first and last 25 uses of the device. [1]The results seen with the COPA in this study thus may parallel initial work on the LMA where, for example, a clear airway was obtained in 95%[7,8]compared with > 99%[3,4]for more recent studies. Performance and anesthetist opinion of the COPA should be expected to improve with increasing individual experience, understanding, and application of the anatomic issues involved.

The variations in performance between the two devices studied here may reflect their different designs, ultimate anatomic positions in the pharynx, and methods of placement. For example, the cuff of the LMA occupies the narrow distal end of the conical laryngopharynx, whereas the cuff of the COPA occupies the wider proximal end of the hypopharynx. The need for airway manipulations with the COPA is probably related to the variable gap between the distal end of the device and the laryngeal inlet. This highlights the importance of selecting the proper size for the given patient, which may, in fact, not always have been the case in this early evaluation of the COPA. A large interlumen gap can become obstructed by the base of the tongue due to the action of gravity [9]combined with reduced tone in the anesthetized pharyngeal muscles. [10]Simple adjustments in head and neck position probably improve airway quality by stretching anterior neck and pharyngeal structures, thus compensating for reduced tone. These maneuvers also elevate the epiglottis, as shown by Morikawa et al. [11]and Boidin, [12]thus improving gas exchange. The interlumen gap with the LMA is negligible, and as long as the epiglottis is not overfolded and entrapped, airway supportive measures are not required. Differences seen in leak pressure and cuff volume between the devices are reasoned by the fact that the LMA cuff occupies the narrower distal end and is more tightly applied to the glottis, whereas the COPA seals the tissues in the more compliant hypopharynx. Although the lower pressure seen with the COPA may limit the frequency and type of pharyngeal complications when compared with the LMA, it may limit its performance in positive-pressure ventilation. This requires further study.

The LMA is inserted more deeply into the pharynx than is the COPA; the former requires positioning of the mask behind the laryngeal opening at the level of C2–3 and the latter requires positioning of the cuff at the base of the tongue at C1–2. This may have led to the greater frequency of blood on the LMA and more frequent sore throat immediately and 1 day after the procedure than with the COPA. A factor that has been shown to reduce the incidence of sore throat with the LMA is cuff pressure control, [13]yet this was not attempted with either device in the present study. Multiple attempts at insertion probably increase the incidence of sore throat with the LMA [14]and was suggested as one of the reasons for sore throat in a previous study with the COPA, [1]although it was not seen in this trial.

We conclude that with respect to physiologic alterations using the devices and overall clinical problems, the COPA and LMA are equivalent. The LMA was associated with a higher first-time placement rate and fewer manipulations during use, suggesting that it is easier to use. The COPA was associated with less blood on the device and sore throat, suggesting that it may cause less pharyngeal trauma. Ultimately, both devices can be used to establish a safe and effective airway for spontaneously breathing anesthetized adults.

The authors thank the data collectors, nurses, surgeons, and the Royal Brisbane Hospital Ethics Committee for their assistance with this study.

*Greenberg RS, Toung T: The cuffed oro-pharyngeal airway-a pilot study. 1992; 77:A558.