The Seoul National University Hospital/Seoul National University College of Medicine Institutional Review Board (Seoul, Republic of Korea) approved this single-centre, parallel-group, randomized controlled trial (reference number: 2217-107-1387; approved 16 January 2023). The trial was registered with CRIS.nih.go.kr (KCT0008151; first submitted 2 February 2023) prior to patient enrolment. All patients provided written informed consent before participating in this trial. The trial was conducted in accordance with the ethical principles of the Helsinki Declaration 2013 and the Good Clinical Practice guidelines, and the manuscript was written in accordance with the applicable Consolidated Standards of Reporting Trials (CONSORT) statements.

We recruited adult (≥ 18 yr old) patients scheduled for elective neurointervention under general anesthesia at Seoul National University Hospital. We excluded patients with upper airway disease (tumour, polyp, trauma, or abscess), cervical spine disease, history of surgical treatment of the upper airway or cervical spine, and high risk of aspiration (gastrointestinal reflux disease), bleeding (coagulopathy), or dental injury (weak or loose teeth).

Before enrolling patients, an anesthesiologist not involved in this study created a random allocation table with computer-generated blocks consisting of six allocations and kept it in an opaque envelope. Just prior to the induction of anesthesia, a nurse not involved in the trial allocated patients to either the posterior-only (wearing only the posterior piece, but not the anterior piece of the cervical collar) or the anterior-posterior (wearing both the anterior and posterior pieces of the cervical collar) group in a 1:1 ratio based on the random allocation table. Patients and anesthesiologists who measured cervical spine angles and investigated intubation-associated complications were blinded to the group allocation. The radiolucent cervical collar used in this study (Philadelphia® Tracheotomy Collar, Össur, Reykjavik, Iceland) also ensured that the anesthesiologist measuring cervical spine angles was blinded to the group allocation.

Prior to the induction of anesthesia, airway-associated variables (modified Mallampati class, interincisal gap, thyromental distance, sternomental distance, neck circumference, and retrognathia) were measured in the sitting position, while thyromental height was measured in the supine position. After preoxygenation with 100% oxygen (6–8 L·min−1) until the fraction of expired oxygen reached 0.8, anesthesia was induced with target-controlled infusion of remifentanil (effect site concentration, 4.0–6.0 mg·L−1) and intravenous bolus injection of propofol (1.0–2.0 mg·kg−1). After loss of consciousness and intravenous bolus injection of rocuronium (0.6–0.8 mg·kg−1), manual mask ventilation was performed with 100% oxygen and sevoflurane (1.5–2.0 vol%). The patient’s head and neck were placed in the neutral position without a pillow.

In the anterior-posterior group, the anterior and posterior pieces of the cervical collar were worn at the front and back of the neck, respectively, and fastened as tightly as possible. In the posterior-only group, only the posterior piece of the cervical collar was worn at the back of the neck without covering the front of the neck with the anterior piece of the cervical collar. To determine cervical spine angles before intubation, a lateral cervical spine radiograph was obtained using the capture method of a biplane angiographic system (Integris Allura™, Philips, Amsterdam, Netherlands).

After confirming the absence of twitches evoked by the train-of-four stimulation, one of two attending anesthesiologists with more than 100 videolaryngoscopic intubations in patients wearing a cervical collar performed videolaryngoscopic intubation. A videolaryngoscope (AceScope™, Ace Medical, Seoul, Republic of Korea) with a disposable Macintosh-style blade (AceBlade™, Ace Medical; size MAC 3 for females and 4 for males) was gently inserted into the oral cavity, avoiding neck extension. After placing the blade tip at the vallecula, the videolaryngoscope was lifted to expose the glottis to a target percentage of glottic opening score of 50% on the monitor. If the glottis was not visible, an assistant performed external laryngeal manipulation, which pressed the thyroid cartilage backward. If the glottis was still not visible, the direct epiglottis elevation method, in which the videolaryngoscope is lifted after placing the blade tip under the epiglottis, was used to expose the glottis. In cases in which the percentage of glottic opening score could not reach 50%, the maximum achievable percentage of glottic opening score was recorded. After then, a tracheal tube (Shiley™ Oral RAE Tracheal Tube with TaperGuard™ Cuff, Covidien, Dublin, Ireland; internal diameter, 7.0 mm for females and 7.5 mm for males) with a malleable stylet bent to 60° at the proximal edge of the endotracheal cuff was advanced until its tip reached the glottic opening. After placing the tube tip at the glottic opening, another lateral cervical spine radiograph was obtained to determine cervical spine angles at intubation. To avoid inconsistent delays in intubation times caused by waiting for the radiologic technologist to take this second lateral cervical spine radiograph, intubation time was defined as the time interval from inserting the videolaryngoscope into the oral cavity to placing the tube tip at the glottic opening just before requesting the radiograph, not to placing the tracheal tube in the trachea. After removing the stylet, the tracheal tube was inserted into the trachea.

The presence of a regular capnogram on the patient monitor confirmed intubation success. If intubation time exceeded 3 min or peripheral oxygen saturation decreased to < 90%, the case was recorded as failed intubation, and rescue manual mask ventilation with 100% oxygen and sevoflurane was performed for ≥ 1 min or until the peripheral oxygen saturation reached 100%. In cases of failed intubation, intubation was attempted again by removing the anterior piece of the cervical collar or using a different intubation device.

At the end of neurointervention, extubation was performed and the patient was transferred to the postanesthesia care unit. During extubation, the anesthesiologist checked for blood in the oral cavity and tracheal tube as well as injuries to the tongue and teeth. Hoarseness and sore throat were assessed in the postanesthesia care unit and on the ward ward at 1 and 24 hr after intubation, respectively. The severity of sore throat was evaluated using a numeric rating scale, with 0 representing no pain and 10 the most severe pain imaginable.

All lateral cervical spine radiographs were saved in the Picture Archiving and Communication System (M6 version 6.0.12.1, IFINITT Healthcare Co. Ltd., Seoul, Republic of Korea). The reference lines of the occiput and C1 were defined as a line connecting the sellar base and the opisthion and that connecting the inferior cortical margin of the C1 anterior arch and the inferior cortical margin of the C1 spinous process, respectively (Fig. 1).13 The reference lines of C2 and C5 were defined as a line connecting the anteroinferior cortical margin of the C2 body and the inferior cortical margin of the C2 spinous process and that parallel to the endplate of the C5 body, respectively.13 An anesthesiologist, who had previously used the identical method to measure the same cervical spine angles in other studies, measured cervical spine angles before and at intubation at the occiput–C1, C1–C2, and C2–C5 segments at the intersections of the aforementioned reference lines.13,14 Each cervical spine angle was measured twice on the same lateral cervical spine radiograph, and the average of the two measurements was used for analysis to reduce measurement error.

Example of reference lines for occiput (A), C1 (B), C2 (C) and C5 (D)

The primary outcome measure was cervical spine motion during intubation, defined as the change in cervical spine angle (calculated as cervical spine angle at intubation minus that before intubation) at the occiput–C1, C1–C2, and C2–C5 segments. The secondary outcome measures were the cervical spine angles before and at intubation at the three aforementioned segments, intubation performance (intubation success rate, intubation time, percentage of glottic opening score, frequency of external laryngeal manipulation, and direct epiglottis elevation method), and intubation-associated complications (incidence of airway bleeding [oral cavity and tracheal tube], airway injury [tongue and teeth], hoarseness, and sore throat as well as severity of sore throat).

In a previous study, the mean (standard deviation) of cervical spine motion during videolaryngoscopic intubation at the occiput–C1 segment was 6.8 (5.0)° in patients wearing both the anterior and posterior piece of the cervical collar.15 Also, another study reported a 4° difference (14° − 10°) in median cervical spine motions during videolaryngoscopic intubation at the same segment between manual in-line stabilization and cervical collar application.16 Based on these findings, we assumed a 3.4° (6.8°/2) difference in cervical spine motions during intubation at the same segment between the posterior-only and anterior-posterior groups. Accordingly, the minimum required sample size was calculated to be 91 when setting the α, β, and effect size to 0.017 (0.05/3), 0.2, and 0.68 (3.4/5.0), respectively. Considering a dropout rate of 10%, we recruited 51 patients per group.

All statistical analyses were performed with the intention-to-treat method using statistical software (IBM SPSS Statistics for Windows version 26, IBM Corp., Armonk, NY, USA). The Shapiro–Wilk test was used to assess the normality of continuous variables. The Student’s t test and Mann–Whitney U test were used to compare normally and nonnormally distributed continuous variables between the posterior-only and anterior-posterior groups, respectively. To calculate confidence intervals (CIs) for difference in means of normally distributed continuous variables between the two groups, the Student’s t test was used. To calculate CIs for difference in medians of nonnormally distributed continuous variables between the two groups, the Hodges–Lehmann estimation was used. The paired t test or Wilcoxon signed-rank test was used to compare cervical spine angles before and at intubation when their distribution was normal and nonnormal, respectively. Fisher's exact test was used to compare categorical variables between the two groups. To calculate CIs for differences in percentages of categorical variables between the two groups, direct computation of the exact CI was used.17 To compensate for multiple comparisons of cervical spine motion and angles at the occiput–C1, C1–C2, and C2–C5 segments between the two groups, a Bonferroni correction was applied and a P value < 0.017 (0.05/3) was considered statistically significant. Otherwise, a P value < 0.05 was considered to be statistically significant.