This was a prospective, crossover, randomized controlled trial of a simulated difficult airway with concurrent limited mouth opening and neck movement. Ethical approval for this study was approved by the institutional review board of Hallym University Kangnam Sacred Heart Hospital (approval number: 2020-09-009). The trial was registered prior to patient enrolment at www.clinicaltrials.gov (NCT04716218) on 20/01/2021, and conducted and reported according to the Consolidating Standards of Reporting Trials (CONSORT) 2010 statement [20]. This study was conducted from January 2021 to March 2021. We recruited 64 adult patients undergoing general anesthesia with tracheal intubation for elective surgery. All patients provided written informed consent and had ASA physical status of I–III. The exclusion criteria were aspiration risk, bleeding tendency, uncontrolled hypertension, obstructive sleep apnea, and a known or anticipated difficult airway (e.g. Mallampati class III–IV, body mass index > 35 kg/m2, inter-incisor distance < 3.5 cm, thyromental distance < 6 cm).
Using a computer-generated randomization table (www.randomizer.org), the enrolled patients were randomly assigned to one of the two sequences. In sequence 1, the back-up position was used first, followed by the neutral position. In sequence 2, the neutral position was used first, followed by the back-up position. The sequence allocation was concealed in an opaque envelope, which was opened by the investigator responsible for randomization immediately before the initiation of anesthesia.
Study protocol
A McGrath™ MAC VL (Aircraft Medical, Ltd., Edinburgh, UK) with a curved blade (resembling a Macintosh blade) and a camera was set up according to the standard practice at our institution. Tracheal intubation was performed using a size 3 McGrath™ MAC VL blade, and a TT with an internal diameter of 6.5 mm for females and 7.5 mm for males (Mallinckrodt, St. Louis, MO, USA). A malleable stylet was not initially used to facilitate tracheal intubation, because the McGrath MAC VL had a Macintosh-geometry blade [7, 21].
All patients were premedicated with glycopyrrolate (0.2 mg) at least 30 min before initiation of anesthesia. After entering the operating room, each patient was placed in the supine position on the operating table, with their head on a soft pillow. Standard monitoring (continuous pulse oximetry, electrocardiography, capnography, and non-invasive blood pressure) was applied.
After removing the pillow, an observer blinded to the sequence assignment assessed the airway based on the Mallampati class, thyromental distance, and inter-incisor distance at maximal mouth opening. Then, a semi-rigid cervical collar (Philadelphia Cervical Collar Co., Thorofare, NJ, USA) of appropriate size was placed around the patient’s neck, and the inter-incisor distance at maximal mouth opening was re-measured. The back-up position was created by flexing the operating table at the trunk–thigh hinge and then raising the back section. For each patient, a digital inclinometer and tripod water level were used to determine the table ramp angle required to align the sternal notch and EAM in the same horizontal plane.
Anesthesia was induced with propofol (1.5 mg/kg) and remifentanil (0.5–1 µg/kg), followed by rocuronium (0.6 mg/kg). After verifying neuromuscular blockade using a nerve stimulator, tracheal intubation was performed using the McGrath MAC VL according to the assigned sequence. In sequence 1, the initial assessment of laryngeal view was performed in the back-up position (end of period 1). Subsequently, the patient was placed in the neutral position and the second assessment of laryngeal view was performed (start of period 2). The trachea was then intubated in the second position. For the sequence 2, the procedure was reversed. Thus, the laryngeal view was assessed in both the neutral and back-up positions, but the trachea was intubated only once (in the second position). When assessing laryngeal visualization, external laryngeal manipulations (ELMs) were not permitted. To eliminate interobserver variation, a single experienced anesthesiologist (> 100 previous tracheal intubations using the McGrath MAC VL) performed all VL procedures.
Outcome measurement
The following data were collected by two investigators (Eun Hee Chun and Joo Hyun Jun) not involved in the VL procedures.
The primary outcome was the laryngeal view in both the neutral and back-up positions (periods 1 and 2, respectively), which was assessed according to the percentage of glottic opening (POGO) score and modified Cormack–Lehane (MCL) grade. The POGO score reflects the proportion of the glottic area that is visible: a score of 100% denotes visualization of the whole glottis, from the inter-arytenoid notch to the anterior commissure, whereas a score of 0% denotes visualization of none of the glottis [22]. Based on the MCL grade, the laryngeal view was classified as easy (laryngeal inlet visible; MCL grade 1 or 2a), restricted (posterior glottic structures or epiglottis visible, where the latter could be lifted; MCL grade 2b or 3a), and difficult (epiglottis could not be lifted or no laryngeal structures visible; MCL grade 3b or 4) [23]. The data collection form included illustrations of the MCL grades and POGO scores to promote standardization.
The secondary outcome was the ease of intubation, which was measured in the second position (period 2). To evaluate this outcome, we recorded the optimization maneuvers used for successful tracheal intubation, such as withdrawing and reinserting the blade, increasing the lifting force, applying ELMs, bending the TT into a steeper curve, adding a stylet, or rotating the TT during passage into the trachea to avoid impacting the anterior wall of the subglottic space [24, 25]. The time to intubation, Intubation Difficulty Scale (IDS) score [26], reasons for failed intubation, and adverse effects were also recorded. Time to intubation represented the time between insertion into and removal of the VL blade from the mouth.
Statistical analysis
Statistical analyses were performed using SAS (version 9.4; SAS Institute, Cary, NC, USA), SPSS (version 27.0; IBM Corp., Armonk, NY, USA), and R software (version 4.1.0; http://www.R-project.org). The distribution of continuous data was evaluated using the Shapiro–Wilk test. Normally distributed continuous variables are provided as the mean ± standard deviation and were analyzed using the paired and independent t-tests. Non-normally distributed continuous variables are provided as the median (interquartile range) and were analyzed using Wilcoxon’s signed rank and Mann–Whitney U tests. Categorical data are expressed as n (%) for proportions and were compared using the McNemar and χ2 tests, with bootstrapping applied as appropriate. To evaluate possible carryover effects, the sum of the POGO scores in the first position (period 1) and second position (period 2) was calculated for each subject and compared across the two sequences using the unpaired t test. To evaluate possible period effects, the difference in POGO scores between the two periods was calculated for each subject and compared across the two sequences using the unpaired t-test [27]. In all analyses, P <0.05 was considered statistically significant.
We calculated the sample size based on the POGO score using PASS software (version 15.0; NCSS, LLC, Kaysville, UT, USA). Assuming that a 20% difference in POGO score was clinically important (standard deviation of 35%) [28], we determined that 28 pairs were required in each sequence for two-sided testing in this 2 × 2 crossover study, with 95% power and an alpha level of 5%. Therefore, we enrolled 32 participants in each sequence group, assuming a dropout rate of 10%.