The Effect Of Different Duration Of One-lung Ventilation On Lung Injury

Since the first video-assisted thoracoscopic surgery (VATS) was reported in the nineties, the use of VATS has become widespread. VATS is now more generally used than the open thoracotomy owing to its minimal invasiveness and the low morbidity. However, due to small view of surgery, VATS requires surgical instruments to complete the operation, thus operation time maybe is longer than the traditional thoracotomy. One-lung ventilation (OLV) is a commonly used technique that facilitates surgical visualization during thoracic surgical procedure. In particular, during VATS, OLV may play a vital role in the successful completion of the surgery. Degrees of damage of OLV on ventilated lung and non-ventilated lung was different.The major impact of OLV on ventilated lung is mechanical stretch-induced lung injury. Mechanical stretch includes:mechanical positive pressure ventilation, lung tissue repeated collapse and reexpansion during OLV, lung tissue stretch by surgical operators. But the impact of OLV on non-ventilated lung is larger than that of ventilated lung. The non-ventilated lung is known to be hypoperfused owing to hypoxic vasoconstriction, pulmonary blood flow decreases by20-25%and the collapsed lung alveoli have no oxygen. So that collapsed lung experiences the process of collapse to re-expansion, just is the same as the process of hypoxia and reoxygenation. It is an ischemia-reperfusion injury. A lot of clinical researches have proved that lung re-expansion can evoke severe oxidative stress, oxygen free radicals was associated with OLV time. The long time of OLV increases postoperative complications. How is the impact of duration of OLV on lung injury, and how long is the long time OLV? Is it3hours,4hours, or longer? Data haved showed that plasma IL-6and IL-8increased both in OLV and two-lung ventilation (TLV), but OLV may increase higher. The concentration of IL-6and IL-8increased in OLV can resume to the similar level as those in TLV after48hours. This demonstrates that once we have a good control of OLV duration, the clinical application of OLV anesthesia is safe. But this study only observed that lung injury was relieved when OLV was only two hour, there is no literature about the impact of longterm OLV on lung injury.Currently, there is no clear data description the time limitation of OLV. Therefore, this study attempts to observe the effect of different duration of OLV on lung injury from clinical and animal experiments, and to find out the time limitation of OLV. We still observe the repair of lung injury in24hours after OLV through animal experiments, and expect to find out the mechanism of lung injury and repair in the future.Objective1. Retrospective analysis of pulmonary complications correlative factors after video-assisted thoracic surgery.2. To search the effect of different duration of OLV on rabbit Tumor necrosis factor-alpha(TNF-α) and Interleukin-8(IL-8) in bronchoalveolar lavage fluid and Surfactant protein A(SPA) expression in lung tissue.3. To search the effect of different duration of OLV on clinical patients’TNF-a and IL-8in bronchoalveolar lavage fluid (BALF).Methods1. The correlative factor analysis of pulmonary complications after video-assisted thoracoscopic surgery1.1Patients330patients (ASA physical status I or II,212male patients,118female patients, age range18-77years, body weight42-80kg) undergoing elective video-assisted thoracic surgery at Guangzhou Institute of Respiratory Disease from2010.1.1to2011.3.31were enrolled in the study. Exclusion criteria included:lung infection or other infectious diseases before surgery; empirical application of anti-infective therapy before surgery; tracheotomy and other traumatic operation and long-term mechanical ventilation; mediastinal thymoma with myasthenia gravis; intraoperative cardiopulmonary resuscitation and emergency treatment; transfer to open-thoracic surgery; a long time to restore two-lung ventilation during operation; preoperative chemotherapy and radiotherapy; stop smoking for less than2weeks before surgery.1.2General anaesthesiaMidazolam0.05mg/kg and atropine0.01mg/kg were intramuscular injected30mins before anesthesia. The size of Broncho-Cath(?) double-lumen tube (DLT) was determined by the tracheal inner diameter at the level of sternoclavicular articulation measured in the chest x-ray. The intubation depth of DLT was determined based on the regressive equation.Patients had the vital signs monitor after getting into the operation room. After inhalation of oxygen for5min, Anaesthesia was induced with target controlled infusing propofol (target blood concentration3ug/ml), sufentanyl0.4μg/kg and cisatracurium0.2mg/kg. A flexible styletted DLT was accomplished via direct laryngoscopy according to a standardized protocol, and a fiberoptic bronchoscope (FOB) was then employed to confirm correct positioning of the tube. After intubation, Radial artery and internal jugular vein were catheterized.All the selected patients had the same ventilation mode. The ventilation parameters setting, including tidal volume, the ratio of inspiratory and expiratory, and respiratory rate were basically the same. The anesthesia method and medication, postoperative analgesia and analgesic drugs were also the same. The operations were performed by the same group of surgeons. All patients were sent to ICU for further care after extubation.1.3Logistic regression analysis for possible factors about postoperative pulmonary complications in VATSPossible factors of postoperative pulmonary complications in VATS included: Age, sex, body mass index(BMI), preoperative pulmonary function rating, the forced expiratory volume in first second(FEV1), forced vital capacity(FVC), FEV1/FVC, maximal voluntary ventilation(MVV), echocardiography, duration of OLV, duration of anesthesia, intraoperative blood loss, intraoperative fluids, laterality of surgery, extent of pulmonary resection(inspection without resection, wedge resection, lobectomy, multi-lobectomy, pneumonectomy), length of ICU stay, duration of chest tube drainage, length of postoperative hospitalization, preoperative and postoperative albumin values, the comorbidities of diabetes mellitus, angiocardiopathy, chronic respiratory disease, and history of thoracic surgery. The above data were collected and collated.In the first3days after VATS, pulmonary exudation increasing, pneumonia, atelectasis or lung compression, pulmonary embolism, pneumoderma of chest wall increasing, pleural effusion, respiratory failure, suction with FOB, and reintubation were considered to be postoperative pulmonary complications in VATS.Length of anaesthesia was defined as the period between induction and extubation. Length of OLV was defined as the period between starting OLV10min before operation and resuming two-lung ventilation. According to the Gerald’s definition of long-term operation, the anaesthesia time was divided into4stages(2h<,2-3h,3-4h,≥4h), and the OLV time was divided into4stages(lh≤,1-2h,2-3h,>3h).1.4Statistical analysesSPSS13.0statistical software package was used for all analyses. Descriptive statistics (x±s) were used to summarize the measurement data, medians were used in categorical data. The independent t-test or Chi-square test was used to compare continuous data or categorical data between no complication group and complication group. Binary logistic regression (Forward LR method) was applied to determine the relationship between variables and postoperative pulmonary complications. P values <0.05were considered to be statistically significant.2. The effect of different duration of one-lung ventilation on rabbit TNF-α and IL-8in BALF and SPA in lung tissue2.1Animals and Experimental DesignSixty-eight New Zealand White rabbits (both males, body weight2.5±0.2kg) were forbidden food and water for one day before experiment. All rabbits were randomised divided into6experimental groups:controlled group (group C) and group O1,O2,O3,O4,O5. Group C (n=10) just received endotracheal intubation and vascular puncture but the spontaneous breath was ramained. Groups01-05respectively received right-lung ventilation for1,2,3,4,5hours (group01and02n=10; group O3n=11, group O4n=12, group O5n=15). Each rabbit resumed two-lung ventilation for30minutes after OLV. Half of rabbits in each group were executed at the end of experiment, the other half rabbits were feed for24hours after resuming spontaneous breath, and then were executed.2.2Experimental operationRabbits were anesthetized with sodium pentobarbital (30mg/kg) intravenously administered through the ear vein and maintained at doses of10mg/h. Each rabbit had a cervical tracheotomy with an inserted endotracheal tube (a diameter of3.0mm self-made cuffed endotracheal tube). After intubation, animals were given vecuronim (0.1mg/kg) per hour and ventilated with a tidal volume of10ml/kg, a rate of35breaths/min, inspiration-expiration (I:E) ratio of1:1, an inspired oxygen concentration (FiO2)100%by use of the animal ventilator. Each rabbit had a carotid or femoral artery puncture to measure heart rate (HR) and median arterial pressure (MAP) by a safedraw transducer blood sampling set. Tracheal tube was put forward into the right-main stem bronchus to build OLV model. The sixth of left intercostal space was opened a small incision (1cm×1cm) to observe the lung collapse. The OLV model was successfully established according to the left lung from expansion to collapse. After OLV model established, right OLV was given a tidal volume of8ml/kg, a rate of40breaths/min, I:E ratio1:1, and a FiO2100%. The effectiveness of left lung collapse was checked by continuous inspection. If the oxygen can’t be maintained during experiments, animals were checked endotracheal tube position and resumed two-lung ventilation for15minutes, then continued OLV. Rabbit’s body temperature was kept at37℃by use of a warm blanket and measured with a mercury thermometer in the anus. Physiologic saline solution was given intravenously10mL·kg-1·h-1. At the end of experiment, half of rabbits in each group were executed by injection of20ml of air; the other half of rabbits were sutured trachea and intercostal space after resuming spontaneous breathing, and were executed after feeding for24hours.2.3Exclusion criteriaAccidental death in the experiment, including death before the scheduled observation time; the recovery rate of BALF less than40%; resuming two-lung ventilation more than five times, et al. The data of exclusion rabbit was removed. Animal experiments were continued until there were10rabbits to complete experiments in each group.2.4Times for variables recording and specimen collection2.4.1Times for variables recordingTO:Before the tracheal intubation;T1:Before the beginning of OLV;T2:Resuming two-lung ventilation for30min after the end of OLV;T3:Feeding for24hours.2.4.2The recorded data of heart rate (HR), mean arterial pressure (MAP) and respiratory mechanicsHR, MAP, PETCO2, VT, breath rate, peak inspiratory airway pressure (Ppeak) of group01-05were recorded at T1and T2. If SBP drops to60mmHg, rabbits were given vasoactive drugs to maintain blood pressure stable. MV and lung compliance were calculated by formula. Analyzed the correlation between HR, MAP, PETCO2with duration of OLV at T2.2.4.3Blood Gas AnalysisBlood samples were obtained by taking0.5ml heparinized blood from the carotid or femoral artery. Blood gas analysis of group01-05was performed at T1and T2by an i-STAT blood gas analyzer. Analyzed the correlation between blood gas values with duration of OLV at T2Before intubation, two animal blood gas values were randomly selected from each group and taken as TO values. Blood gas analysis of group C and group01-05were performed at T3. The blood gas analysis was compared between at TO and T3.2.2.4Pulonary histopathological examinationLeft and right lung were taken out at the end of the experiment.1.0cm3lung tissues were taken at the same parts of lower lobes in left and right lung. Lung tissue was soaked in10%buffered formalin and fixed, then embedded in paraffin, slice, row, hematoxylin-eosin stainng. Lung injury score was viewed10high-power (400times) under an ordinary optical microscope. The contents of lung injury score were atelectais, hyaline membrane formation, infiltration or aggregation of neutorphils in airspaces and alveolar congestion or hemorrhage. The injury score was used a damage scale of0to4(0, minimal or no damage;1, mild damage, less than25%of the field;2, moderate damage,25%to50%of the field;3, severe damage,50%to75%of the field;4, maximal damage, more than75%of the field). The total score of the four conetents was calculated as lung injury score.2.4.5Western Blot Analysis for SPA in lung tissue1.0cm3lung tissues were taken at the same parts of lower lobes in left and right lung and were stored in-70℃refrigerator. Western blot was used to determination of SPA expression in lung tissue. The protein was formed image by ChemiDocXRS (BIO-RAD) and the immunoreactive bands were scanned and normalized by GAPDH bands of the same membrane. The band density was analyzed by Image J software. Each sample was repeated3times.2.4.6The concentration of TNF-a and IL-8in BALFThe right lung was inserted a diameter of3.0mm endo-tracheal tube into the right main bronchus and infused5ml of ice-cold sterile PBS. Lavages were repeated three times and the recovery fluid was combined. The left main airway was clamped during right lung lavaged. The left lung was also taken the same method to get BALF after finishing right lung lavage. Supernatant aliquots was used to test the concentrations of TNF-a and IL-8by enzyme-linked immunosorbent assay (ELISA). The correlation between TNF-a, IL-8and duration of OLV were analyzed at T2.2.4.7BALF Cell CountsThe cell pellets of BAL were washed by PBS and added10%acetic acid to remove red blood cells,1ml PBS were taken out and counted under light microscopy.2.5Statistical analysisThe Statistical Package for the Social Sciences13.0(SPSS Inc.; Chicago) was used for results analysis. The results are presented as means±standard error of the mean. The paired-samples t test was used to compare hemodynamics, respiratory mechanism and blood gas values at T1and T2, and compare the concentration of TNF-a, IL-8, cell number, SPA gray-value, lung injury scorce in left and right lung. One-way analysis of variance was used to compare blood gas values at TO and T3. The independent samples t test was used to compare lung injury score, the concentration of TNF-a and IL-8, cell number, SPA gray-value at T2and T3. Analysis of variance of repeated measures data was used to compare different group variables at different time point. Least-significant difference was used for multiple comparisons. The count data was compared by Chi-square test. The correlation between circulation, blood gas values, the concentrations of TNF-a and IL-8with duration of OLV was analyzed by Spearman analysis. P values<0.05were considered to be statistically significant.3. Clinical research:the effect of different duration of one-lung ventilation on patient’s TNF-a and IL-8in BALF3.1Patients groups and selected standardsPatient’s age was range between25-64year; ASA physical status I or II. Patients were assigned into two groups:1) one-lung ventilation group (OLV group, n=36): patients who undergone elective esophageal and lung cancer surgery via left VATS were enrolled in the study. According to OLV time, patients were divided into three subgroups:group01(<1.5hOLV); group02(2.5-3hOLV); group O3(3.0-3.5hOLV),12patients in each group.2) Two-lung ventilation group (T group,3.0-3.5hTLV, n=12):patients who undergone non-thoracoscopic surgery were enrolled in the study.Selected case criteria:no dysfunction of heart, liver and kidney; No preoperative chemotherapy and radiotherapy; no history of mechanical ventilation; Non-small cell lung cancer; the tumor grade T1N1M0(tumor size<3cm, no distant metastasis); no infection history within a week; no history of using corticosteroids and antibiotic; normal lung function or mild pulmonary dysfunction; the recovery rate of BALF was more than40%; re-expand lung less than5times during operation; patients whose blood loss more than15%of weight or transfusion were exclude from study. The patients enrolled in group T had normal pulmonary function and had no lung cancer or endocrine cancer. The case was excluded when the patients need to inhale FiO2>70%more than1hours. The enrolled patients of group T had normal pulmonary function and had no lung cancer or endocrine cancer.3.2General anaesthesia and ventilation methodAnaesthesia was induced and maintained with target controlled infusing propofol. In the OLV group, a flexible styletted DLT was accomplished via direct laryngoscopy according to a standardized protocol, and a fiberoptic bronchoscope (FOB) was then employed to confirm correct positioning of the tube. All patients were placed central venous catheter and radial arterial catheter. Intravenous infusion speed and fluid dosage were decided according to blood loss, urine output and central venous pressure (CVP).The patients were ventilated with a FiO2of0.6through out. Ventilator settings were adjusted from VT of8-10ml/kg, rate of12-min-1(TLV) to VT of7-8ml/kg, rate of12-16-min-1(OLV); I:E ratio was1:2. Ventilator settings were kept constant during the study. Patients were sent to the PACU. All patients’post-operative analgesia and analgesic drugs were the same, analgesic dose was calculated by body weight. Patients were pulled out the endotracheal tube after recovery of spontaneous breathing.3.3Observations time pointsT1:before anesthesia;T2:20minutes after two-lung ventilation in supine positionT3:at the end of surgery (resuming two-lung ventilation for30min in group OLV)T4:in the first postoperative day while the patients were receiving oxygen via nasal catheter (2L/min)3.4Hemodynamics and respiratory mechanicsSpO2, MAP, HR, CVP, Ppeak, Pplateau, MV, VT, PETCO2of group01-03and group T were recorded at T2and T3. SpO2, MAP, HR were recorded at T1and T4.3.5The blood gas analysis values Arterial blood gas analysis was performed at T1to T4respectedly. PH, PaCO2, PaO2, SaO2, oxygenation index (PaO2/FiO2), and bicarbonate were measured by Stat Profile(?) Critical Care Xpress Analyzer.3.6The concentration of TNF-α and IL-8in BALFBALF was performed at T2and T3. Put the top of bronchoscopy to the opening of segment or sub-segmental bronchi, and used silicone tube rapidly injecting sterile normal saline30ml through the biopsyhole (10ml/time, total30ml). Lavage fluid was stored at a clean sputum collection through a vacuum suction and recorded the recovery volume. The supernatant was taken to measure the concentrations of TNF-a and IL-8by ELISA. In group OLV, lavage bilateral pulmonary, first preferred lavage lung segment in the left lower lobe and right lower lobe, second preferred lavage lung segment in the left upper lobe and right upper lobe. In group T, we only lavage the right middle lobe segment and taken as controlled value.3.7Other observation items(1) duration of OLV;(2) duration of anesthesia;(3) the amount of blood loss and urine;(4) intraoperative intravenous input;(5) the residence time in the PACU;(6) the amount of narcotic drugs;(7) record the abnormal chest X-ray results in3days after surgery in OLV group;(8) record complication occurrence in3days after surgery in group OLV, such as hypoxemia, pneumonia, reoperation incidence, pulmonary atelectasis, etc.3.8Statistical analysesDescriptive statistics (x±s) were used to summarize the continuous data, medians were used in categorical data. SPSS13.0statistical software package was used for all analyses. One-way analysis of variance was used to compare patients’ general situation and measurement data. Least-significant difference was used for multiple comparisons. Chi-square was used to compare count data. One-way analysis of variance was used to compare the hemodynamic values and blood gas values at different time point in the same group, and the paired-samples t test was used to compare mechanics values and TNF-α, IL-8values at different time point in same group. Analysis of variance of repeated measures data was used to compare hemodynamics, respiratory mechanics, blood gas values and the concentrations of TNF-α, IL-8at same time point in defferent groups. P values<0.05were considered to be statistically significant.Results1. The correlative factor analysis of pulmonary complications after video-assisted thoracoscopic surgery1.1Postoperative Pulmonary Complications (PPC)72of300patients (21.88%) suffered PPC in the first3days after VATS.24pulmonary exudation increasing (7.27%), pneumonia in23(6.97%),18atelectasis (5.45%),4pleural effusion (1.21%),2suction with FOB (0.61%),1reintubation (0.3%). No mortality occurred.1.2Compared with non-PPC group, Preoperative albumin was lower(P<0.05), length of OLV and ICU stay were longer than in PPC group (P<0.01).1.3Binary Logistic Regression (Forward LR)Preoperative serum albumin level, length of OLV, length of ICU stay were independent risk factors, the values of OR were0.922,1.379and1.362(P<0.05).2. The effect of different duration of one-lung ventilation on TNF-a, IL-8in BALF and SPA in lung tissue2.1One rabbit restored two-lung ventilation for7times in group05; there had1,2and4rabbit respectively in group03,04and05were excluded during experiment. There had no difference in body weight, Vt and MV in each group (P>0.05)2.2The comparison of HR and MAP before and after OLVThe repeated measures ANOVA analysis showed that HR and MAP had no significant difference between groups (HR F=1.142, P=0.349; MAP F=2.363, P=0.067). HR and MAP at different time point had significant difference (HR F=69.699, P=0.000; MAP F=23.546, P=0.000); HR of each group had no interaction with time (F=1.996, P=0.111), but MAP of each group had interaction with time (F=6.612, P=0.000), these suggested that HR had the same trend but MAP has not.Further analysis of MAP, the results showed that the comparison of group01 and02, group04and05had significant difference (P<0.05), MAP of each group had no difference at T1(P=0.867).2.3The comparison of PETCO2, Peak, lung compliance before and after OLVThe repeated measures ANOVA analysis showed that PETCO2, Peak, lung compliance had no significant difference between groups (PETCO2F=0.119, P=0.975; Ppeak F=0.258, P=0.903; Cdyn F=0.435, P=0.782). The comparison of PETCO2, Peak, lung compliance had significant difference at different time point (PETCO2F=22.575, P=0.000; Ppeak F=56.687, P=0.000; Cdyn F=58.345, P=0.000). PETCO2, Peak, and lung compliance of each group had interaction with time (P<0.037), these suggested that the trend of PETCO2, Peak, Cdyn is different with time.Further analysis, the results showed that the comparison of PETCO2in group04and05had significant difference (P<0.05), Peak, and lung compliance in group03to05had significant difference (P<0.05); PETCO2, Peak, and lung compliance had no significant difference at T1and T2in each group(P>0.172).2.4Blood gas analysis in each groupThe repeated measures ANOVA analysis showed that pH, PaCO2and bicarbonate had significant difference between groups (pH F=2.770, P=0.038; PaCO2F=3.098, P=0.025; bicarbonate F=2.609, P=0.048), PaO2、BE、SaO2had no significant difference (PaO2F=0.592, P=0.670; BE F=1.450, P=0.233; SaO2F=1.568, P=0.199); pH, PaCO2,PaO2, bicarbonate and SaO2had significant difference (P=0.000). PH, PaO2, bicarbonate and SaO2had no interaction with time (P>0.124), PaCO2of each group had interaction with time (P=0.004), these suggested that the trend of pH, PaO2, bicarbonate and SaO2is same with time, but the trend of PaC02is different with time.Further analysis, the results showed that the comparison of PaCO2in group04and05had significant difference between T1and T2(P<0.01); PaCO2of each group had no significant difference at T1(P=0.740), PaCO2of each group had no significant difference at T2(P=0.001).2.5The comparison of blood gas values between24h after OLV and before anesthesia At T3, ph and PaCO2of group04-05, PaO2and SaO2of group01-05, Bicarbonate and BE of group05had difference with those at TO and in group C (P<0.05)Bicarbonate of group C had difference with that in TO (P<0.01), the other observations had no statistical difference (P>0.05)2.6The correlation of blood circulation, blood gas analysis with OLV time at T2HR was positively correlated with OLV time (HR r=0.396, P<0.01); MAP was negatively correlated with OLV time (MAPr=-0.537, P<0.01). PETCO2, Ppeak and Cdyn had no correlation with OLV time (P>0.05)pH, PaO2, BE and SaO2were negatively correlated with OLV time (pH r=-0.451, PaO2r=-0.373, BE r=-0.502, SaO2r=-0.421, P<0.01), PaCO2and bicarbonate were positively correlated with OLV time (PaCO2r=0.466, bicarbonate r=0.412,P<0.01).2.7Pathological changes and lung injury score2.7.1SpecimensThere had no significant changes in left and right lung in group C and01. But in group O3-O5, the colour of left lung was dim and the surface was congestion, the right lung also had surface congestion in some part of the lobes. The group C and group01at24h later had no difference with those at OLV; different degrees of congestion were still in group02-05in24hours later, especially when OLV time was longer than3h.2.7.2Light microscope examinationDuring OLV: In group C and group01, alveolar was integrity, alveolar spaces had no exudation and less inflammatory cells. But in group03-05left lung, pulmonary interstitial was obvious thicken, pulmonary interstitial and alveolar space had inflammatory cells and blood cell infiltration, alveolar wall was destructed and alveolar atelectasis, especially in group05. In the right lung of group03-05, pulmonary interstitial was obvious thichening and alveolar space was less exudation than in left lung. 24h after OLV: there had no obvious exudation in group C and O1, and a little atelectasis was seen in group O1. In group O2, there had alveolar collapse, intertitial congestion thichen, and red blood cells in left lung; there also had neutrophil in right lung alveolar. In group O3, alveolar collapsed, interstitial thichening and red blood cells in alveolar spaces. In group O4, alveolar was obvious collapsed and exudation, many red blood cells and neutrophils were seen. In group O5, lung collapsed obviously and a large number of red blood cells and macrophage infiltration were seen.2.7.3Lung tissue injury scoreThe repeated measures ANOVA analysis showed that lung injury score had significant difference at T2and T3(P=0.000); lung injury score also had significant difference between left and right lung(T2F=6.748, P=0.012; T3F=7.779, P=0.007). lung injury score of each group had interaction with time at T2(F=5.121, P=0.001), but it had no interaction with time at T3(F=0.262, P=0.932), these suggested that the trend of the lung injury score is different with time at T2, but the trend is same with time at T3.Further analysis, the results showed that the comparison of lung injury score in group05had significant difference at T2(F=6.748, P=0.012); lung injury score of each group had significant difference between left and right lung (left lung F=69.303, P=0.000; right lung F=19.909, P=0.000)Left lung injury score of group O5was higher at T2than at T3(P>0.05).2.8Western blot for SPA expression in lung tissue2.8.1The repeated measures ANOVA analysis showed that SPA of left lung had significant difference at T2and T3(F=9.540, P=0.001); SPA had significant difference in different time point (F=21.884, P=0.001); SPA of each group had interaction with time (F=14.652, P=0.000), these suggested that the trend of SPA is different with time. Further analysis, the results showed that the comparison of SPA in group O4and O5had significant difference between T2and T3(P<0.001), there had significant difference between groups at different time point (P<0.05) 2.8.2The repeated measures ANOVA analysis showed that SPA of right lung had significant difference at T2and T3(F=10.990, P=0.000); SPA had significant difference in different time point (F=359.691,P=0.000); SPA of each group had interaction with time (F=14.171, P=0.000), these suggested that the trend of right lung SPA is different with time. Further analysis, the results showed that the comparison of SPA in group01to05had significant difference between T2and T3(P<0.05), there also had significant difference between groups at different time point (P<0.01)2.8.3The repeated measures ANOVA analysis showed that SPA of left and right lung had significant difference at T2(F=10.876, P=0.000); SPA had significant difference between left and right lung (F=36.394, P=0.000); SPA of each group had interaction with left and right lung (F=11.577, P=0.000), these suggested that the trend of each group is different with left and right lung. Further analysis, the results showed that the comparison of SPA in group01to O3had significant difference between left and right lung (P<0.05), there also had significant difference between left and right lung in different groups (P<0.01)2.8.4The repeated measures ANOVA analysis showed that SPA of left and right lung had significant difference at T3(F=19.860, P=0.000); SPA had significant difference between left and right lung (F=0.083, P=0.778); SPA of left and right lung had no interaction with groups(F=1.946, P=0.1600), these suggested that the trend of left and right lung is same with different groups.2.9The comparison of TNF-α、IL-8in BALF2.9.1The repeated measures ANOVA analysis showed that TNF-a of left and right lung had significant difference at T2and T3(T2F=56.146, P=0.002; T3F=14.619, P=0.001); TNF-a had significant difference between groups at T2(F=29.336, P=0.000), but TNF-a had no significant difference between groups at T3(F=1.054, P=0.410). TNF-a of each group had interaction with left and right lung at T2(F=4.792, P=0.005), but TNF-a of each group had no interaction with left and right lung at T3(F=0.333, P=0.888), these suggested that the trend of TNF-a in each group is different with left and right lung at T2, but the trend of TNF-a in each group is same with left and right lung at T3.Further analysis, the results showed that the comparison of TNF-a in group03to05had significant difference between left and right lung at T2(P<0.05), there also had significant difference between left and right lung in different groups (P=0.000).TNF-a of left BALF in group03-05was higher at T2than at T3(P<0.05).2.9.2The repeated measures ANOVA analysis showed that IL-8of left and right lung had significant difference at T2and T3(T2F=9.460, P=0.005; T3F=11.784, P=0.002); IL-8had significant difference between groups at T2(F=20.703, P=0.000), but IL-8had no significant difference between groups at T3(F=5.824, P=0.001) IL-8of each group had interaction with left and right lung at T2(F=2.882, P=0.035), but IL-8of each group had no interaction with left and right lung at T3(F=2.452, P=0.063), these suggested that the trend of IL-8in each group is different with left and right lung at T2, but the trend of IL-8in each group is same with left and right lung at T3.Further analysis, the results showed that the comparison of IL-8in group04had significant difference between left and right lung at T2(P=0.043), there also had significant difference between left and right lung in different groups (P<0.01)IL-8of left BALF in group04-05was higher at T2than at T3(P<0.01) IL-8of right BALF in group05was higher(P<0.05).2.10The correlation of TNF-a and IL-8values and duration of OLVAt T2, TNF-a and IL-8of left and right lung were positively correlated with duration of OLV (left lung TNF-a r=0.901, right lung TNF-a r=0.748, left lung IL-8r=0.840, right lung IL-8r=0.660, P<0.01). At T3, TNF-a in left BALF and IL-8in left and right BALF were positively correlated with duration of OLV(left lung TNF-a r=0.430, P<0.05; left lung IL-8r=0.697, right lung IL-8r=0.592, P<0.01). 2.11BALF cell countsThe repeated measures ANOVA analysis showed that cell counts of left and right lung had significant difference at T2and T3(T2F=13.606, P=0.001; T3F=47.137, P=0.000); cell counts had significant difference betw...