This study protocol was approved (April 2012) by our local Research Ethics Board (Comité d’éthique à la recherche de l’Institut universitaire de cardiologie et de pneumologie de Québec - Université Laval). This was a prospective randomized single-blinded (i.e., only the patients and the thoracic surgeons were blinded to the group assignment) controlled study in patients requiring one-lung ventilation (OLV) for VATS. Written informed consent was obtained from consecutively scheduled adult patients undergoing OLV for elective left or right VATS. Exclusion criteria included patients with previous or anticipated difficult intubation, prior chemotherapy or thoracic radiotherapy, severe chronic obstructive pulmonary disease or asthma (forced expiratory volume in one second < 50% predicted), and pleural and/or interstitial pathology. There were some a priori post-randomization exclusions, including bronchoscopic findings precluding the use of the specific randomized devices (e.g., origin of the right upper lobe (RUL) bronchus is too close to the carina, precluding use of the BB) and severe oxygen desaturation occurring at any time during the observation period (i.e., until complete lung collapse was achieved) necessitating ventilation of the non-dependent lung.
Before anesthesia induction, patients were assigned by computer-generated randomization to one of two study groups, left-sided DL-ETT (Mallinckrodt™ left endobronchial tube; Mallinckrodt Medical, Cornamaddy, Athlone, Westmeath, Ireland) or BB (Fuji Uniblocker; Fuji Systems, Tokyo, Japan) with the internal channel intentionally occluded to exclude any passive contribution to lung collapse. Just prior to anesthesia induction -and after the anesthesiologist confirmed the absence of a potentially difficult airway- research staff opened a sealed envelope indicating the device to be used. Thoracic surgeons were blinded to group assignment throughout the procedure.
Anesthesia management included application of standard monitoring15 and a standardized intravenous induction of anesthesia with propofol and sufentanil, with muscle relaxation (rocuronium) used according to local practice. Anesthesia was maintained with an inhaled anesthetic (sevoflurane or desflurane). For patients with a baseline heart rate < 100 beats·min−1, intravenous glycopyrrolate 0.2 mg was administered prior to induction to reduce the production of tracheobronchial secretions.
Lung isolation methods
After induction of anesthesia, the patient’s trachea was intubated with either a single-lumen 8.0-mm internal diameter endotracheal tube (SL-ETT) for the BB group or a left-sided DL-ETT (37 Fr for females, 39 Fr for males). All BBs were placed via the SL-ETT using a flexible bronchoscope (FOB). Bronchial blockers were positioned distally in the main bronchus with their cuff deflated. Left-sided DL-ETTs were positioned with the FOB at an appropriate depth to allow positioning the tracheal carina between the radiopaque black line of the endobronchial aspect of the DL-ETT and the upper part of the blue bronchial cuff.16 An attending anesthesiologist or a resident directly supervised by an attending anesthesiologist (who confirmed all device placements) placed the devices for both the BB group and the left-sided DL-ETT group.
Prior to OLV, mechanical ventilation operated in a volume-controlled mode with a tidal volume of 8-10 mL·kg−1 (ideal body weight), respiratory rate of 10 breaths·min−1, positive end-expiratory pressure (PEEP) of 5 cm H2O, and F
i
O2 of 1.0. After the patients were positioned in a lateral decubitus position, the BB cuff or DL-ETT bronchial cuff was inflated under FOB guidance, and OLV was initiated using volume-controlled ventilation (tidal volume of 5-7 mL·kg−1 ideal body weight with a respiratory rate of 14-16 breaths·min−1, PEEP of 5 cm H2O, and F
i
O2 of 1.0). The internal channel of the BB was deliberately occluded. After the pleura had been open for 20 min, the F
i
O2 was adjusted to maintain O2 saturation > 95%.
In the left-sided DL-ETT group, lung isolation was initiated by clamping the bronchial or tracheal lumen of the Y-connector (corresponding to the lung to be isolated) and opening the corresponding lumen of the DL-ETT to allow for passive lung deflation. Just before inflating the BB balloon in the BB group, a 30-sec apnea period was instituted immediately prior to initiating OLV. A second 30-sec apnea period, also with the BB balloon transiently deflated, was initiated at the time of the pleural incision. In the left-sided DL-ETT group, a “sham” apnea period was performed at the beginning of OLV to ensure study blinding of the thoracic surgeon. For the sham apnea period in the DL-ETT group, the anesthesiologist made similar gestures as in the BB group but without opening the airway or stopping the ventilation. A second “sham” apnea period was also conducted in the DL-ETT group at the time of pleural incision.
Surgeons were absent from the operating room during DL-ETT or BB placement and blinded to the airway device by means of a drape placed over the lung isolation device and the endotracheal tube. The FOB video monitor was oriented such that the surgical team could not identify the lung isolation device being used.
To assess the time of lung collapse and objectively evaluate the quality of lung deflation, we established a scoring system generally based on previously published studies.4-8 Fig. 1 outlines the standardized definition of lung collapse on a three-point visual and descriptive scale, where 1 = no lung collapse; 2 = partial lung collapse; and 3 = total lung collapse. The thoracic surgeons and the three observers performing the offline video examinations of lung deflation also used this three-point scale.
The primary endpoint was the time from when the surgeon opened the pleura until complete lung collapse (i.e., score of 3 on the three-point scale) as determined from the offline analysis of video recordings taken during the procedure. For the video analysis, a DVD recorder (Sony RDRHX730, Sony Corporation, Malaysia) linked to the surgical camera (Olympus CV-180, Olympus Canada Inc., Richmond Hill, ON, Canada) was used to record the entire duration of surgery. The VOB files obtained were converted to WMV files and edited with Windows Movie Maker (Microsoft Corporation; Microsoft Canada Headquarters, Mississauga, ON, Canada). This allowed the three observers blinded to group assignment (i.e., two thoracic surgeons and one cardiothoracic anesthesia fellow) to analyze the video examination data offline (i.e., 30-sec clips edited every five minutes from the videos recorded during VATS). The video clips reviewed were mixed from random group and time sequences and scored using the same standardized three-point scale as previously described. Several secondary endpoints were also recorded. The offline quality of lung deflation was evaluated at standardized time points -i.e., immediately on pleural opening (0 min) and five, ten, and 20 min after pleural opening, using the same visual and chart scale as for the primary endpoint.
In addition to performing the offline assessments, the thoracic surgeons conducted a clinical evaluation of the time to achieve complete lung collapse and the quality of lung deflation using the same scale as for the offline evaluation. Other secondary endpoints included having the surgeon guess the type of device being used. This was done 20 min after pleural opening and before any exploration of the hilum. The use of any suction to assist lung collapse was also recorded.
Statistical analysis
All data were analyzed using the statistical package program SAS® 9.3 (SAS Institute Inc., Cary, NC, USA). Our sample size estimate was based on the time to complete lung collapse evaluated by video examination (i.e., our primary endpoint). We used data from Campos et al.6 where they reported a mean (standard deviation [SD]) collapse time of 18 (7.2) min after initiation of OLV in the DL-ETT group. Assuming a 40% difference between groups with an alpha of 0.05 and a power of 0.80, we estimated that we would require a sample size of 18 patients per group. We randomized 40 patients (20 subjects per group) to account for potential post-randomization exclusions.
Clinical observation data were log-transformed prior to undergoing analysis of variance (ANOVA). Statistical results from these parameters were expressed with transformed values. Tukey’s multiple comparison technique was applied post hoc to the ANOVA to compare pairs of group means. The univariate normality assumption was verified using the graphical representations and the Shapiro-Wilk test. The Brown-Forsythe variation of Levene’s test statistic was used to verify the homogeneity of variances. The data regarding time to lung collapse were analyzed using a mixed model with heterogeneous variances for the two techniques. Quality of lung collapse was analyzed with Fisher’s exact test. All reported P values are two sided. As the secondary outcome of quality of lung collapse was evaluated at four different times (0, 5, 10, and 20 min after lung isolation), a Bonferroni correction was applied and P < 0.0125 was considered significant.
Video examination data were also secondarily analyzed using a model based on the probability of observing a given lung deflation score at a given time. From the model, the parameter T50,3 (time associated with 50% probability of observing a complete lung collapse score of 3) was estimated using the solver function in Microsoft Excel. The Appendix includes a description and analysis of the model. The magnitude of interobserver agreement regarding the video examination data was determined using a weighted Kappa statistic where 0-0.20 would denote slight agreement; 0.21-0.40, fair agreement; 0.41-0.60, moderate agreement; 0.61-0.80, substantial agreement, and > 0.81 almost perfect agreement.17