In a prospective, randomised study design, adult New Zealand white rabbits were obtained as surplus stock from the Animal Resource Center of the University of Calgary and a commercial supplier (Charles River Laboratories, Québec city, QC, Canada, n = 12 males, n = 3 females). Rabbits from the commercial supplier had a minimum of 1 week to acclimatize before experimentation. Rabbits were either group or individually housed (depending on location) in an environmentally controlled unit [12 h light/12 h dark cycle (lights on at 0730), target temperature between 16 and 22°C, target humidity between 40 and 50%]. All rabbits were provided with ad libitum pellets, hay, and water with a small amount of fruit and vegetables. Experiments were performed between 1700 and 2000 hours.

Animals were block randomized to one of the two treatment groups for managing airway and each device type was placed by a single investigator: ETT (Daniel S. J. Pang) SGAD (Sarah Engbers). These investigators both completed an online training program provided by the manufacturer and practical experience was gained with two rabbits scheduled for euthanasia. Investigators could not be blinded to group assignment, with the exception of the veterinary pathologist (Cameron Knight), who performed necropsies off-site. Animals were fasted for approximately 2 h before the experiment. Physical examinations (cardiothoracic auscultation, examination of musculoskeletal system, and integument) were performed immediately before each experiment. Exclusion criteria included any indication of systemic illness resulting in assigning an American Society of Anesthesiologists physical classification status >2.

Animals were sedated with dexmedetomidine (0.1 mg kg⁻¹, Dex-domitor, Zoetis, Kirkland, QC, Canada) and midazolam (0.5 mg kg⁻¹, Midazolam Sandoz, Sandoz Canada, Boucherville, QC, Canada) delivered intramuscularly as a single injection in the epaxial musculature. Ten minutes later, animals were placed in sternal recumbency on the CT scan table, provided oxygen by facemask (1.5 L min⁻¹) and cannulas (22 gauge 1″, BD Insyte, Becton Dickinson Infusion therapy systems Inc., Sandy, UT, USA) placed in the auricular vein (IV) and artery. A baseline CT scan (Toshiba Aquilon 8 slice, Markham, ON, Canada) was performed under sedation following instrumentation, followed by a second scan after ETT or SGAD placement. Scans were performed by a board-certified radiologist (Nicolas Rousset), with the scan margins ranging from the rostral margin of the nose to the thoracic inlet.

Following the baseline CT scan, invasive systemic arterial blood pressure (BP), ECG, and saturation of arterial hemoglobin with oxygen (pulse oximetry probe on the prepuce or vulva) monitoring began (LifeWindow LW6000V, Digicare Animal Health, Boynton Beach, FL, USA). General anesthesia was induced with alfaxalone (0.3 mg kg⁻¹ IV, Alfaxan, Jurox, Rutherford, NSW, Australia). The SGAD size was chosen based on body mass as per the manufacturer’s recommendations. Where an animal fell between sizes, the larger SGAD was selected. An insertion attempt began 30 s after completing alfaxalone injection. Each animal in the SGAD group was positioned in sternal recumbency with its tongue exteriorized and the lubricated SGAD (lubricating spray, Docsinnovent Ltd., London, UK) inserted into the oropharynx until either resistance was encountered, or the incisors were within 1–2 cm of the device’s fixation tabs. Acceptable placement was confirmed with a mainstream capnograph attached to the SGAD (Masimo EMMA Capnograph with pediatric sample chamber, Masimo, Danderyd, Sweden) and visualization of six rectangular waveforms. Criteria for rescue intubation with ETT for the SGAD group were defined as: lingual cyanosis that did not resolve with repositioning the SGAD or tongue, or three unsuccessful insertion attempts (no capnograph trace observed).

Orotracheal intubations were performed using a blind technique (19). Endotracheal tube size (2–3.5 mm ID, Sheridan/CF, Teleflex Medical Canada Inc., Markham, ON, Canada) was selected based on investigator experience. Each animal was positioned in sternal recumbency with the head and neck hyperextended. A single dose of lidocaine (12 mg lidocaine per metered dose, Lidodan, Odan Laboratories Ltd., Montreal, QC, Canada) was sprayed into the caudal oropharyngeal area 30 s before attempting ETT. Confirmation of ETT was with a mainstream capnograph, as described for SGAD placement. For both SGAD and ETT, the number of insertion attempts and total time for insertion were recorded. Insertion time was from the moment the SGAD or endotracheal tube passed the incisors until the first of six capnograph waveforms was observed. If capnograph waveforms were not observed following SGAD placement, adjustments were made by repositioning slightly rostrally or caudally. The number of attempts during intubation with each device was recorded, with an “attempt” representing the withdrawal of the device out of the mouth before re-inserting or changing to a different size. Airway devices were secured behind the ears with a bandage tie.

Baseline systemic arterial BPs were recorded immediately before induction of general anesthesia. Beginning immediately after the insertion of the airway device, physiologic data (BP, heart, and respiratory rates, SpO₂) were collected and recorded every 30 s for 5 min, and at 15 min post-insertion. Additionally, the highest systemic systolic arterial BP observed during device insertion was recorded (from video-recording of the physiologic monitor). At 10 min, the second CT scan was performed. At 15 min, the concentration of isoflurane measured from a sampling line placed approximately 5 cm from the mouth was recorded, and the airway seal was tested by closing the adjustable pressure limiting valve on the anesthetic breathing system and squeezing the reservoir bag. Peak inspiratory pressures were increased in 5 cmH₂O increments, between 5 and 20 cmH₂O, until a leak was heard (or 20 cmH₂O achieved). This leak test was done by the same investigators (Sarah Engbers and Daniel S. J. Pang) for all animals. Occurrences of lingual cyanosis or airway obstruction (audible wheeze during respiratory cycle or irregular capnograph trace) were recorded. In the last nine animals (SGAD n = 4, ETT n = 5), an arterial blood sample was collected from the intra-arterial cannula 30 min after device placement and analyzed for blood gases and electrolytes (CG8+ cartridges, iSTAT Portable Clinical Analyzer, Heska Corporation, Fort Collins, CO, USA). General anesthesia was maintained for 60 min (from time of induction) with isoflurane (1% vaporizer setting) carried in oxygen (1 L min⁻¹) through a non-rebreathing system (coaxial Bain). All animals were euthanized with an overdose of IV sodium pentobarbital while anesthetized at the end of the experiment.

CT scans were reconstructed using a multi-planar reconstruction tool on OsiriX software (v 7.0.1, Pixmeo, Geneva, Switzerland).

All measurements were taken three times by a board-certified veterinary radiologist (Nicolas Rousset) and averaged for analysis. The midline sagittal view of each SGAD rabbit post-device placement was used to measure the following distances: (1) distance between the caudal edge of the basihyoid bone (BH) to the rostral edge of the thyroid cartilage (TC) (A–B, Figure 1), (2) the distance from the caudal edge of the BH to the tip of the SGAD (A–C, Figure 1), and (3) the distance between the SGAD tip and the rostral edge of the TC (B–C, Figure 1). Confirmation of whether the SGAD tip entered the upper esophageal sphincter was recorded in each instance, as well as the anatomical location of the SGAD tip. Transverse images from before and after device placement were compared to determine the narrowest region of the upper airway. In the SGAD group, the site of the narrowest section was also measured relative to the SGAD tip in order to determine any contribution from the device to airway narrowing. The cross-sectional area (CSA) of the narrowest point was measured. For the ETT group, the luminal CSA was measured.

Necropsies were performed on all rabbits within 2 h of euthanasia by a board-certified veterinary pathologist (Cameron Knight) according to a standardized method (20). The tongue, pharynx, larynx, and trachea were collected and fixed in 10% neutral-buffered formalin. Tissues were prepared routinely for histologic evaluation and stained with hematoxylin and eosin. For each rabbit, three to four representative transverse sections from each of the following locations were examined histologically: (1) larynx at the level of the vocal folds, (2) trachea immediately caudal to the larynx, and (3) trachea 1 cm caudal to the larynx. The mucosal and submucosal layers in each section were evaluated according to a scoring system modified from a previously published study (7). The modification involved removing scoring criteria that reflected chronic injury from repeated endotracheal intubation over several weeks, which were not applicable in the current study. The scoring criteria used in this study are listed in Table 1. Scores for sections from the same region were averaged and a total score (/6) was calculated by summing the mucosal and submucosal scores.

Commercial software was used for statistical analyses (Prism v 7.0a, GraphPad Software, La Jolla, CA, USA). Data were tested for normal distribution with a D’Agostino and Pearson normality test. Non-parametric analyses were used if data were not normally distributed or sample sizes were too small for the normality test. Mann–Whitney tests were used to make the following comparisons between groups: body mass, anatomical distances, cross-sectional airway areas, device insertion time, airway seal, percentage changes in systolic BP during device insertion, arterial blood sample parameters, and histological scores. A Wilcoxon matched pairs test was used to compare changes in systolic arterial BP over time. Two-way ANOVA was applied to systemic arterial BPs, heart rate, respiratory rate, and expired CO₂ data, followed by a Bonferroni post hoc test if main effects were significant. A p value < 0.05 was considered significant. Data are presented as mean ± SD and median (range). 95% confidence intervals (95% CI) for mean or median differences are provided as available.

Two rabbits were excluded from the ETT group, one due to failure to achieve orotracheal intubation and the other due to prolonged intubation resulting from fecal matter in the oropharynx (21). CT scans were not performed in two rabbits (both from the ETT group) due to unavailability of the CT scanner. In these cases, the anesthetic duration was matched to the full procedure and other data were collected as planned. All rabbits in the SGAD group had a device placed. Final group sizes were: ETT, n = 6; SGAD, n = 7. Rabbits ranged in age from 3 to 12 months. There were no significant differences in body mass between groups: ETT; 2.6 (2.3–4.5) kg, SGAD; 2.7 (2.4–5.0) kg (p = 0.38, 95% CI −2.3 to 1.8). All rabbits appeared healthy on physical examination and were assigned an American Society of Anesthesiologists physical classification status of 1. The endotracheal tubes used in the ETT group ranged from 2 to 3.5 mm ID: 2 mm, n = 2; 2.5 mm, n = 1; 3 mm, n = 1; 3.5 mm, n = 2. Three sizes of SGAD were used: R3, n = 4; R4, n = 1; and R5, n = 2.

Airway anatomy was similar between groups. The distance between the caudal edge of the BH and the rostral edge of the TC did not differ significantly (p = 0.16, 95% CI −6.7 to 1.3): SGAD, 16.2 (13.2–20.3) mm; ETT, 14.5 (12.7–15.8) mm (distance A–B, Figure 1).

In most cases (6/7 rabbits), placement of the SGAD resulted in the tip of the device deviating ventrally to be positioned within the laryngeal vestibule (n = 6, Figures 2 and 3). This resulted in a reduction in the CSA of the airway. The CSA before device placement was smallest within the nasopharynx in nine rabbits and within the larynx in two rabbits, with no significant difference between groups (p = 0.78, 95% CI −2.3 to 3.7 mm²) so that baseline CSA was pooled. The CSA of the narrowest point decreased significantly between the baseline CT scan [CSA, 12.2 (6.9–13.4) mm²] and placement of the SGAD [CSA, 2.7 (2.0–12.3) mm², p = 0.003, 95% CI 4.1 to 10.3] and ETT [CSA, 3.8 (2.3–6.6) mm², p = 0.002, 95% CI 4.6–10.6, Figure 4]. There was no significant difference in CSA between ETT and SGAD (p = 0.93, 95% CI −5.7 to 3.9). The site of the smallest CSA in the SGAD group was located within the device tip [4.2 mm (range 3.2–5.2) from the tip extremity] at an average of 5.3 mm (range 2.9–7.9) caudal to the rostral edge of the TC, so that the smallest CSA was formed by the SGAD dorsally and the rabbit’s airway ventrally (Data Sheet S1 in Supplementary Material). In the ETT group, a uniform CSA was created by the endotracheal tube.

Device insertion time was shorter and more consistent (narrower range) for the SGAD [33 (14–38) s] than ETT [59 (29–171) s, p = 0.02, 95% CI 12–133, Figure 5]. The number of attempted device placements did not differ between groups (SGAD: 1 attempt, n = 4; 2 attempts, n = 3. ETT: 1 attempt, n = 4; 2 attempts, n = 2. p > 0.99, 95% CI −1 to 1). The median airway seal was higher in the ETT group [15 (10–20) cmH₂O], but there was no statistical difference compared to the SGAD group [5 (5–20) cmH₂O, p = 0.06, 95% CI 0–15, Figure 6]. A single animal in the SGAD group had an airway seal of 20 cmH₂O. In this animal, loss of the capnograph trace indicated respiratory obstruction, so the SGAD was retracted until the trace returned and the airway seal test performed after repositioning. The median SGAD insertion depth was 20.7 (11.1–26.5) mm (distance A–C, Figure 1), and the rabbit in which the SGAD was adjusted had the shortest insertion distance (11.1 mm), so that the device tip did not enter the larynx. Isoflurane was not detected, nor lingual cyanosis observed, in any rabbit. Lingual cyanosis was not observed in any rabbit of either treatment group.

The ETT group was found to have significantly more damage to tracheal mucosa and submucosa [3.3 (1.0–5.0)] than the SGAD group [0.67 (0.33–3.67), p = 0.03, 95% CI 0–3.3, Figure 7].

There was a systematic, albeit non-significant, difference in systemic arterial BP between groups, with higher BPs observed in the SGAD group (systolic, p = 0.45; diastolic, p = 0.11; mean, p = 0.17; Data Sheet S2 in Supplementary Material). The percentage change in peak systolic arterial BP observed during placement did not differ between each group and baseline (SGAD, p = 0.3, 95% CI −5.8 to 3.9; ETT, p = 0.16, 95% CI −13.2 to 1.1). Between groups, ETT [3.2 (−1.1 to 13.2)%] showed a wider range of percentage change in peak systolic BP than SGAD [1.0 (−3.9 to 5.8)%] though the median difference was not significant (p = 0.45, 95% CI −11.9 to 2.1, Figure 8). There were no significant main effects between treatment groups for heart rate (p = 0.32), respiration rate (p = 0.22), and expired CO₂ (p = 0.74), though heart and respiratory rates tended to be higher in the SGAD group (Data Sheet S2 in Supplementary Material). One rabbit in each group had a single SpO₂ reading <90% (SGAD, 89%; ETT, 88%; both were recorded immediately following device insertion).

No significant differences in the following variables were observed: pH, PaO₂, PaCO₂, BE, HCO₃, SaO₂, Na, K, iCa, hema-tocrit, hemoglobin (Table S1 in Supplementary Material). Hypercapnea was present in both treatment groups: SGAD; 65.1 (54.0–75.1) mmHg, ETT; 51.8 (41.2–67.8) mmHg.