Lung isolation
Lung isolation was performed before opening of the pleura. The conditions of lung isolation were as follows: the pressure inside the sleeve capsule of the left tube was low, between 15 and 20 cm; the volume inside the sleeve capsule was 2-3 ml; and auscultation of isolated lung respiratory tone disappeared. Because arterial blood pressure and heart rate might increase temporarily after anesthesia induction, the tracheal intubation was implemented at 5 min after anesthesia induction with 0.1 mg/kg imidazole, 0.05 mg/kg midazolam, 1.5–2 μg/kg sufentanil and 0.1 mg/kg vecuronium. DLT was inserted into patients of Group D, and the tube position was confirmed by flexible bronchoscopy (FOB) (Olympus BF-240 bronchoscope; Olympus America, Pennsylvania, USA) in left semi-lateral position at 45 degree. Meanwhile, single lumen tube combined with a BB (Tappa Medical Technology Co., Ltd., Hangzhou, China) was inserted into patients of Group B, and positioned under FOB guidance in the operative position. Notably, tube displacement was detected by fiberoptic bronchoscopy after left semi-lateral position at 45 degree, and the best position of the tube was adjusted by means of fibronchoscopy.
A 7.5–8 # tracheal tube (Mallinckrodt™ Lo-Contour Oral/Nasal T real Tube Cuffed, Mallinckrodt Hi-Lo (TM), Covidien, Dublin, Ireland) or a 37 #/39 # left DLT (Mallinckrodt™ Endobronchial Tube) was selected in male patients, while a 7–7.5 # tracheal tube or 35 #/37 #/39 # left DLT was used in female patients. In group D, 35 #, 37 #, and 39 # DLTs were used in 10, 11, and 9 patients, respectively. In Group B, 7 #, 7.5 #, and 8 # single lumen tubes were used in 9, 16, and 4 patients, respectively.
Transesophageal echocardiography was performed after completing intubation, and mechanical ventilation was set in the volume-controlled ventilation model at 8–10 ml/kg of ideal body weight with 95% oxygen concentration. Once the ascending aorta was blocked, oxygen supply was stopped, and the tracheal tube cuff was deflated. Oxygen supply was regained by one-lung ventilation after removing the blockage. When the SpO2 became < 90% (PaCO2 < 60 mmHg), the cuff of the BB in Group B was partially relaxed, and the right lung was expanded to meet the gas exchange requirement. When removing the BB after the surgery, the tracheal tube was pulled out as appropriate. In Group D, double lung ventilation was performed until SpO2 was > 90% (PaCO2 > 60 mmHg). The surgery was performed under single lung ventilation. After the surgery, the DLT in Group D was replaced with a single-lumen tracheal tube using a tube exchanger or a direct laryngoscop.
MICS
CPB was maintained after lung isolation. For mitral valve surgery, the aorta was clamped, cardiac arrest was induced by perfusion of cardioplegia, and the oxygen supply was stopped. When valve replacement was completed, the left atrium was closed, the clamped aorta was released, and the oxygen supply returned. Aortic blockage was not need for tricuspid valve surgery. Before closing of the chest, the patient was weaned off CPB, and double lung ventilation was performed.
Outcome measurements
The primary outcome was arterial oxygen partial pressure (PaO2). Other outcomes were also collected, including: 1) the times of intubation and tube localization; 2) the incidences of tube displacement (> 1.5 cm was defined as a significant displacement), 3) postoperative sore throat, and hoarseness; and 4) the mean arterial pressure (MAP) and heart rate (HR) before and after intubation; 5) the time of cardiopulmonary bypass (CPB); 6) maintenance time (T) for SpO2 < 90% (PaCO2 < 60 mmHg); 7) peripheral oxygen saturation (SpO2), PaCO2, EtCO2, and mean airway pressure and airway peak pressure before (T1) and 30 min after (T2) one-lung ventilation, 30 min after CPB (T3), and 30 min before (T4) the end of the operation; and 8) surgeons’ satisfaction with anesthesia that were evaluated by the same surgical surgeon.
Patients were followed-up for 3 months after the surgery, and the occurrence of short-term complications were recorded, such as peritoneal leakage, artificial valve thrombosis, thromboembolism, acute respiratory failure, hepatic insufficiency, low cardiac output syndrome, and multiple organ failure.
Statistical analysis
SPSS 18.0 software (SPSS Inc., Chicago, IL, USA) was used for statistical analysis. The sample size was conducted using Power analysis and sample size (PASS) software package program (Utah, USA, version 11.0 for windows). Continuous variables were expressed as mean ± standard deviation (SD), and categorical variables as percentages. Differences between different groups were compared using t-test for continuous variables, and chi-square (χ2)-test for categorical variables.