Patient enrollment and randomization: This study was approved by the institutional ethical committee of Guangzhou Women and Children’s Medical Center (No. 2014051229, approval date: June 3, 2014). The trial was registered prior to patient enrollment at China Clinical Trial Registry (http://www.chictr.org.cn/showproj.aspx?proj=4344, Principal investigator: Yingyi Xu, Registration number: ChiCTR-TRC-14005232, Date of registration: 12 August 2014). Written informed consent was obtained from all patients enrolled in the study. Informed consent was signed by the guardians of each patient. Pediatric patients who would undergo elective thoracic surgery between September 2014 and June 2016 at Guangzhou Women and Children’s Medical Center were selected. The enrollment criteria were as follows: (1) ASA stage I-III and (2) age of 0.5-3 years old. The exclusive criteria were as follows: (1) airway compression; (2) laryngeal edema or acute airway inflammation; (3) abnormal takeoff of the right upper lung lobe; (4) foreseeable difficulties in endotracheal intubation. The enrolled patients were randomized into the bronchoscopy (BRO) group and the CT group using the closed envelope technique. The endobronchial blocker placement was guided by bronchoscopy in the BRO group and by CT 3-dimentional airway reconstruction in the CT group. All cases of anesthesia were performed by a pediatric anesthetist with 6-year experience of thoracic anesthesia. Random numbers were generated using software (SAS 9.2, SAS Institute Inc, Cary, NC, USA) with a ratio of 1:1. These numbers were then sealed in envelopes and kept by an independent study coordinator who did not participate in anaesthesia, perioperative care and postoperative follow-up of the patients. During the study period, patients were consecutively recruited and randomly divided into the control or intervention group accordingly. Anaesthesiologists who gave anaesthesia did not participate patients’ follow up and data collection. Patients, healthcare providers and investigators who were in charge of follow-up and data collection, were blinded to the study protocol.
CT measurement: All pediatric patients received cervical and chest CT scanning (5 mm thickness, 5 mm interval, Aquilion 64, Toshiba) under sedation at the supine position for 3-dimentional airway reconstruction before surgery. The FH plane was determined with the bilateral auriculares and the orbitale; the median vertical plane was determined with the middle of sella turcica, the nasospinale, and the posterior edge of foramen magnum. The distance from the incisor teeth to the carina was measured at the vertical plane when the airway between the incisor teeth and the carina was clearly exposed ; if the patient was not at a proper position or the airway was compressed, the distance would be measured after surface reconstruction (Figure 1).
Anesthesia: All pediatric patients received intravenous injection of 0.01 mg/kg penehyclidine hydrochloride before surgery and oxygen inhalation after entering the operation room. They received micro-pump infusion (8-10 ml/kg/h) of sodium acetate Ringer's injection, and their BP, HR, ECG, and SpO2 were monitored. Midazolam (0.05 mg/kg), sufentanil (0.3 μg/kg), and rocuronium (0.6 mg/kg) were injected intravenously to induce general anesthesia. Then, an tracheal tube without side holes (Weili Medical Inc, Guangzhou, China) was intubated under direct vision of laryngoscopy. The catheter model was selected according to the calculation using the classic formula (based on predicted age formula). After intubation, the partial pressure of carbon dioxide in endexpiratory gas (PETCO2) as well as invasive arterial blood pressure and central venous pressure were monitored, and tracheal aspiration was performed. Inhalation of 1%-3% sevoflurane was used for anesthesia maintenance with a tidal volume of 6-8 ml/kg. The concentration of sevoflurane was adjusted according to hemodynamic changes and data of anesthesia monitoring. Rocuronium and sufentanil were supplemented while necessary. All patients were subjected to ICU care after surgery.
Endobronchial blocker placement: In the BRO group, the insertion depth of tracheal tube was calculated using the classic formula 12. After the 5 French (5F) Weili endobronchial blocker (Weili medical Inc, Guangzhou, Guangdong, China) was placed into the tracheal tube, an electrobronchoscope (A20-2.8, Maidehao Co, Zhuhai, Guangdong, China) with a diameter of 2.8 mm was inserted to help locate the endobronchial blocker until the point A of endobronchial blocker reached the take off of the main bronchus at the blocking side (Figure 2). The proper blocker placement was confirmed under bronchoscopy after the patients shifted from the horizontal position to the lateral position.
In the CT group, CT 3-dimentional reconstruction images were used to measure the length of the main bronchus (the length from the incisor teeth to the carina) before endobronchial blocker placement. Before endotracheal intubation, the insertion depth was preset as the CT-measured length of the main bronchus minus 2 cm and was marked (marker 1) on the tracheal tube (Figure 3a). The endobronchial blocker was inserted through the tracheal tube until the point A of the sacculus reached the catheter tip. The positions on the blocker which paralleled the screw cap (marker 2) and the screw cap plus 2 cm (marker 3) were marked, then the endobronchial blocker was extubated after the cap was screwed up (Figure 3b). The tracheal tube was inserted to marker 1 under direct-vision laryngoscopy. The endobronchial blocker was inserted through the tracheal tube again. The connectors of the endobronchial blocker and the tracheal tube were fixed when the screw cap paralleled marker 2. The endobronchial blocker was further inserted until the screw cap paralleled marker 3 with resistance disappeared, and the sacculus was inflated with 1.5-2.5 ml of air (Figure 3c). Both lungs were auscultated to make sure that respiratory sounds disappeared in the lung of the blocking side. If proper blocking was not achieved after 5 consecutive repositionings, bronchoscopy-guided placement would be applied, and the patient would be excluded. The proper blocker placement was confirmed by auscultation after the patients shifted from the horizontal position to the lateral position.
Observational parameters: (1) the required time for successful blocker placement (measured since the endobronchial blocker was inserted through the vocal cord until it was placed at the proper position); (2) the number of repositionings for successful blocker placement (each extubation of the endobronchial blocker from the Tracheal tube was counted as one repositioning); (3) the successful rate of the first blocker positioning; (4) the degree of lung collapse ranked by the surgeon as excellent (complete lung collapse at the blocking side), fair (lung collapse at the blocking side with a little amount of residual air that would not affect surgical exposure), moderate (partial lung collapse which requires suction or manual collapse), and poor (no collapse of the lung) 13; (5) airway mucosal injury graded using bronchoscopy after surgery by an anesthetist as none (no mucosal edema), mild (mild mucosal edema), moderate (obvious mucosal edema and hyperemia), severe (mucosal erosion and hemorrhage) 11; (6) pulmonary infection occurred within 72 h after surgery, which was defined as plaque-like shadow on both lungs with or without pleural effusion observed by chest X-ray; (7) trachyphonia after tracheal extubation; (8) the duration of postoperative mechanical ventilation; (9) the duration of postoperative intensive care unit (ICU) stay; (10) the duration of postoperative hospitalization.
Estimation of sample size: The sample size was estimated with α = 0.05 and 1-β = 0.8 using the PASS 15.0 software (NCSS, Utah, USA). According to our previous clinical experience, the adequacy of lung collapse was similar in the two groups. According to the estimation, at least 61 patients in each group needed to be enrolled to find a moderate variation (i.e., W = 0.3) between the two groups.
Statistical analyses: The SPSS 15.0 software (NCSS, Utah, USA) was used for statistical analyses. Continuous data with normal distribution are expressed as mean ± standard deviation and were analyzed using the independent-sample t test; continuous data with abnormal distribution are expressed as median (interquartile range) and were analyzed using the Wilcoxon rank-sum test; categorical data are expressed as cases (%) and were analyzed using the Pearson χ2 test or the χ2 test with correction for continuity. P < 0.05 was considered significant.