Study protocol Effect of High-Flow Nasal Cannula vs Conventional Oxygen Therapy for Patients with Thoracoscopic Lobectomy after Extubation Objective: To investigate whether high-flow nasal cannula (HFNC) oxygen therapy is superior to conventional oxygen therapy for reducing hypoxemia and postoperative pulmonary complications (PPC) in patients with thoracoscopic lobectomy after extubation.
Methods: Patients with intermediate to high risk for PPC who underwent thoracoscopic lobectomy were enrolled in this prospective study. Subjects were randomly assigned to either HFNC group (HFNCG) or conventional oxygen therapy group (COG) following extubation. Arterial blood samples were collected after extubation at the following time points: 1, 2, 6, 12, 24, 48 and 72h. Patients with postoperative hypoxemia were measured and PPC or reintubation for respiratory failure were recorded. Adverse events related to oxygen therapy were also documented.
Study population:Consecutive sampling was used to recruit the patients underwent planned thoracoscopic lobectomy because of lung tumor. Patients with intermediate to high risk for postoperative pulmonary complications (PPC) were eligible for participation. To identify such patients, the Assess Respiratory Risk in Surgical Patients in Catalonia (ARISCAT) score was used (Supplement 1). The ARISCAT score≥26 is associated with an intermediate to high risk for PPC [5,6]. Patients were excluded from the study if they were immunocompromised; were pregnant; converted to an open thoracotomy because of poor visualization or bleeding; were aged <18 or >80 years or if informed consent could not be obtained.
Randomization, intervention and weaning protocol:All the patients eligible were transferred to ICU for postoperative monitoring and ventilator weaning at the end of the thoracoscopic lobectomy procedure. Patients were classified into two groups by random figure table following extubation. A random number sequence was generated with STATA statistical software version 12.1. HFNC oxygen therapy group (HFNCG) received a flow rate of 35 to 60 L/min and FiO2 was titrated (from 45% to 100%) by the treating clinician to maintain a peripheral oxygen saturation (SpO2) of 95% or more. The conventional oxygen therapy group (COG) received oxygen via either nasal prongs or facemask with oxygen flow titrated (from 45% to 100%) by the bedside clinician to maintain a SpO2 of 95% or more. HFNC oxygen therapy was delivered by the Optiflow™ system (Fisher & Paykel Healthcare Ltd, Auckland, New Zealand) using a MR850 heated humidifier and a RT202 breathing circuit. Natural air includes about 21% oxygen. If a patient is wearing a nasal cannula or a simple facemask, each additional liter/min of oxygen adds about 4 percentage points for the first 3 liters and only 3 percentage point for every liter thereafter to their FiO2.
Follow the guideline of Difficult Airway Society Extubation Guidelines Group  , the patients were ready for scheduled extubation after tolerating a spontaneous breathing trial in ICU. The decision to extubate was at the discretion of the treating doctors in ICU and no mandatory extubation variables were set.
Clinical assessment and outcomes:Baseline assessment included the evaluation of age, gender, body mass index (BMI), acute physiology and chronic health evaluation (APACHE)Ⅱ score, ARISCAT score, baseline PaO2, arterial oxygen tension to inspiratory oxygen fraction ratio (PaO2/FiO2), oxygen saturation to FiO2 ratio (SaO2/FiO2) and PaCO2 before operation. Asthma, chronic obstructive pulmonary disease (COPD) and smoke history were also recorded. Lung function before operation was measured as well and the value of functional residual capacity (FRC) and forced expiratory volume in one second (FEV1)/ forced vital capacity (FVC) were recorded.
The incidence of hypoxemia (defined as PaO2/FiO2 of 300mmHg or less ) was recorded in the first 72h after extubation and the differences of PaO2, PaO2/FiO2, SaO2/FiO2 and PaCO2 between the two groups were compared. Secondly, the rates of PPC like suspected pneumonia (patient receives antibiotics and meets at least one of the following criteria: new or changed sputum, new or changed lung opacities on chest X-ray when clinically indicated, tympanic temperature >38.3°C, white blood cell (WBC) count >12*10^9/L in the absence of other infectious focus) and atelectasis (opacification of the lung with shift of the mediastinum, hemidiaphragm toward the affected area and compensatory overinflation in the adjacent nonatelectatic lung)  were also documented. Acute hypoxemic respiratory failure was defined by one of the hypoxemic criteria (SpO2 <92 % while breathing at least 10 L/min oxygen, PaO2 <60 mmHg on air or PaO2 <80 mmHg while breathing any supplemental oxygen) and at least one of the following: severe respiratory distress with dyspnoea, accessory muscle recruitment and paradoxical abdominal or thoracic motion, respiratory rate >25 breaths/min, respiratory acidosis with pH <7.30 and arterial carbon dioxide partial pressure (PaCO2) >50 mmHg . Once patients after extubation was found with acute hypoxemic respiratory failure, noninvasive ventilation (NIV) (Bipap Vision with humidification, RESPIRONICS INC, USA) was adopted. If the symptoms of respiratory distress did not improve within 2 hours, then reintubation might be considered. The incidence of NIV requirement and reintubation were also compared.
Adverse effects related to HFNC application and oxygen therapy (air leak, throat or nasal pain and abdominal distension) were also recorded. As the previous studies indicated that it was with high incidence of PPC within 72h following thoracoscopic lobectomy [1,9], the arterial blood gases were consecutively collected and checked at 1, 2, 6, 12, 24 ,48 and 72h after extubation.
Statistical Analysis: Review of data from the three study centers over a 3-year period (2012～2014) revealed about 30% of patients with hypoxemia who underwent thoracoscopic lobectomy after extubation. A sample size of 117 for each group provided 80% power to detect a reduction in hypoxemia from 30% to 15% (alpha = 0.05).
Statistical analysis was performed using SPSS version 19.0. Data were initially assessed for normality and were subject to log-transformation where appropriate . Data between the HFNCG and COG were compared using Chi-square test for equal proportion or Fisher exact test where numbers were small with results presented as the number and percentage. Continuous variables with normal distribution were compared using Student t test and presented as means (standard deviations), whereas skewed data was compared using Wilcoxon rank-sum test and reported as medians (interquartile range). Two-way analysis of variance (ANOVA) for repeated measures with Bonferroni post hoc analysis was used for analysis of the modification of variables over time in the two groups. All analysis was performed on an intention-to-treat basis and a two-sided P＜0.05 was considered to be statistically significant. Figures were drawn using Graphpad prism version 6.0.80>60>92>18>