This single-center, prospective, randomized controlled, non-inferiority clinical trial was approved by the Institutional Review Board of Ethics Committee on Biomedical Research at West China Hospital of Sichuan University and registered at Chictr.org.cn (Trial number: ChiCTR2200064831; Date of registration: 19/10/2022). This study was performed in adhere to the applicable guideline: the consolidated standards of reporting trials (CONSORT). Written informed consent was obtained from all the participants before surgery.

The inclusion criteria were as follows: aged 18–70 years old, American Society of Anesthesiologists (ASA) physical status 1 or 2, scheduled to undergo elective right thoracoscopic lung surgery and requiring a left lateral position. Patients with body mass index (BMI) > 30 kg/m2, anticipated difficult airway, limited neck motion, intraluminal lesions in the left main bronchus, an anatomical problem in the tracheobronchial tree, cardiopulmonary impairment, and patients at potential risk of reflux aspiration were excluded from the study.

The patients were allocated to either the left lateral position group (group L) or the supine position group (group S) in a 1:1 ratio through a random number table generated by a computer. The random allocation sequence was placed in a sequentially coded, sealed, and opaque envelope, which was opened by the anesthetist before anesthesia induction in the operating room. While the study was blinded to data collection personnel, it was not blinded to patients, surgeons, nurses and anesthetists responsible for treating patients.

Standard monitoring protocols in thoracic surgery were followed. All patients received continuously monitoring, including saturation of pulse oxygen (SpO2), electrocardiogram (ECG), non-invasive or invasive blood pressure, capnography, neuromuscular and Bispectral Index (BIS) monitoring. Before induction of anesthesia, patients assigned to group S were positioned in a supine position, while those in group L were positioned laterally with right arm extended to facilitate mask ventilation and intubation (Supplementary file 1). All patients received preoxygenation for 10 min before induction. Anesthesia was induced intravenously with propofol 1.5–2.0 mg kg −1, sufentanyl 0.2–0.3 µg kg −1, and rocuronium 1 mg kg −1. A skilled anesthetist with over 10 years experience in anesthesia for thoracic surgery inserted a left-sided VDLT (NORGAS, Jiangxi Norgas Medical Ltd., Jiangxi, China) through the vocal cords under the Macintosh laryngoscope.After removing the stylet, the VDLT was advanced and rotated 90° clockwise until its appropriate position was showed on the VDLT’s screen. Left side 32 or 35 Fr VDLTs were used in female patients and 35 or 37 Fr VDLTs in male patients. General anesthesia was maintained with either volatile (sevoflurane or desflurane) or propofol (targeting a BIS value between 40 and 60) and remifentanil with sufentanyl boluses as required. Lung protective ventilation was used in all patients’ ventilation management.

The primary outcome was the intubation time. Time to intubation was defined as the time from the insertion of larygoscopic blade into the mouth until the satisfactory placement in the endobronchial lumen achieved confirmed through the video of VDLT. Two time points were recorded, as follows: T1, glottis identification time, was defined as the time from the larygoscopic blade passing between the patient’s lips to identification of glottis; T2, railroading time, was defined as the time from identification of glottis to confirmation of bronchial tube positioning.

The secondary outcomes included: (1) VDLT displacement rate: Fig. 1A depicts the endotracheal tube in its appropriate location. When bronchial cuff’s edge was completely concealed at the main bronchus’ entry and the cuff could not be seen on the camera screen (Fig. 1B) or when the cuff of the bronchus herniating into the carina and the carina could not be seen (Fig. 1C), the VDLT displacement was detected; (2) intubation failure rate: intubation was deemed failure when it could not be accomplished within 150 s or SpO2 dropped to 92% or below. The rescue method would be performed if a tracheal intubation failure happens. When the first intubation attempt failed, the anesthetist immediately performed mask ventilation. The patients would be changed from lateral to supine position if the second intubation attempt failed; (3) the satisfaction of surgeon and nurse: at the end of surgery, the surgeon and nurse were questioned the overall satisfaction of lung isolation and patient position. Satisfaction was ranked as follows: 0 = very dissatisfied, 1 = dissatisfied, 2 = somewhat satisfied, 3 = moderate satisfied, 4 = highly satisfied; (4) intubation-related adverse events: the events included mucosal trauma, lip or dental injuries, esophageal intubation, hoarseness and sore throat 30 min and 24 h after extubation. Sore throat was graded as mild (pain with swallowing), moderate (persistent pain and increasing with swallowing), and severe (pain interfering with eating and require analgesic medication).

The different locations of the endotracheal tube are showed in the photograph. A. The appropriate location of the endotracheal tube. B. The bronchial cuff’s edge was totally hidden at the main bronchus entry’ and the cuff could not be seen on the screen. C. Bronchial cuff herniating into the carina and the carina could not be seen on the screen

Sample size calculation was based on the hypothesis that intubation time in group L is not inferior to that in group S. Using a non-inferiority margin of 10 s, based on the findings of previous research [18, 19], the minimum sample size was calculated to be 80 patients to achieve a power of 80% and a 2.5% risk of a type I error (one-side test). Non-inferiority would have been declared if the upper limit of the 97.5% confidence interval (CI) of the mean difference (VDLT intubation time in group L minus that in group S) in the intubation time was below 10 s.

All data were recorded in a Microsoft Excel database and all statistical analyses were performed using SPSS Statistics 26 (IBM SPSS Statistics for Windows, IBM Corp, Armonk, NY, USA). Kolmogorov–Smirnov test was used to test for normality of distribution. Continuous variables are presented as mean and standard deviation or median and interquartile range. Normally distributed continuous data comparisons were performed using independent t-test. Nonparametric analysis for non-normally distributed continuous variables was performed using the Mann–Whitney U test. Categorical data were shown as number and percentage and analyzed using Chi-square test or Fisher’s exact test. Non-inferiority was assessed for the primary outcome by one-sided 97.5%CI of absolute difference using the independent t-test. The results were considered significant with P < 0.05.