Share this article on:

Choosing a Lung Isolation Device for Thoracic Surgery: A Randomized Trial of Three Bronchial Blockers Versus Double-Lumen Tubes

Narayanaswamy, Manu MBBS, FANZCA*; McRae, Karen MDCM, FRCPC†; Slinger, Peter MD, FRCPC‡; Dugas, Geoffrey MD, FRCPC§; Kanellakos, George W. MD, FRCPC†; Roscoe, Andy FRCA∥; Lacroix, Melanie MD, FRCPC¶

doi: 10.1213/ane.0b013e3181999339
Cardiovascular Anesthesiology: Research Reports

BACKGROUND: There is no consensus on the best technique for lung isolation for thoracic surgery. In this study, we compared the clinical performance of three bronchial blockers (BBs) available in North America with left-sided double-lumen tubes (DLTs) for lung isolation in patients undergoing left-sided thoracic surgery.

METHODS: One hundred four patients undergoing left-sided thoracotomy or video-assisted thoracoscopic surgery were randomly assigned to one of the four lung isolation groups (n = 26/group). Lung isolation was with an Arndt® wire-guided BB (Cook® Critical Care, Bloomington, IN), a Cohen Flexi-tip® BB (Cook Critical Care) or a Fuji Uni-blocker® (Fuji Systems, Tokyo) or with a left-sided DLT (Mallinckrodt Medical, Cornamadde, Athlone, Westmeath, Ireland). Anesthetic management and lung isolation were performed according to a standardized protocol. Each group was randomly subdivided into two subgroups (n = 13/subgroup): immediate suction (at the time of insertion of the lung isolation device) (Subgroup I) or delayed suction (20 min after insertion of the lung separation device) (Subgroup D) according to when suction was applied to the BB suction channel or the bronchial lumen of the DLT. Using a verbal analog scale, lung collapse was assessed by the surgeons, who were blinded to the lung isolation technique.

RESULTS: There was no difference among the lung isolation devices in lung collapse scores at 0 (P = 0.66), 10 (P = 0.78), or 20 min (P = 0.51) after pleural opening. The time to initial lung isolation was less for DLTs (93 ± 62 s) than BBs (203 ± 132) (P = 0.0001). There were no differences among the BBs in the time to lung isolation (P = 0.78). There were significantly more repositions after initial placement of the lung isolation device with BBs (35 incidents) than with DLTs (two incidents) (P = 0.009). The Arndt BB required repositioning more frequently (16 incidents) than the Cohen BB (8) or the Fuji BB (11) (P = 0.032).

CONCLUSIONS: The three BBs provided equivalent surgical exposure to left-sided DLTs during left-sided open or video-assisted thoracoscopic surgery thoracic procedures. BBs required longer to position and required intraoperative repositioning more often. The Arndt BB needed to be repositioned more often than the other BBs.

IMPLICATIONS: This study compared the clinical performance of three bronchial blockers (BBs) with left-sided double-lumen endobronchial tubes in patients undergoing left-sided thoracotomy or video-assisted thoracoscopic surgery. We found that the surgical exposure was similar among the BBs and double-lumen endobronchial tubes but BBs required longer to position and required more intraoperative repositionings.

From the *Gosford Hospital, Gosford, New South Wales, Australia; †Toronto General Hospital, Toronto, Canada; ‡Department of Anethesia, University of Toronto, Toronto General Hospital, Toronto, Canada; §Credit Valley Hospital, Mississauga, Ontario, Canada; ∥University Hospital of South Manchester, Manchester, UK; and ¶Hotel-Dieu de Levis, Quebec, Canada.

Accepted for publication October 1, 2008.

The bronchial blockers were provided free of cost to the institution (Toronto General Hospital, University Health Network, Toronto, Canada).

Address correspondence and reprint requests to Manu Narayanaswamy, MBBS, FANZCA, Gosford Hospital, Gosford, New South Wales 2250, Australia. Address e-mail to

One-lung ventilation (OLV) is often performed during thoracic surgery to facilitate surgical exposure. The most common techniques for achieving OLV are a double-lumen endobronchial tube (DLT) or bronchial blockers (BBs). Within the past decade, three new independent BB have become available (Fig. 1).1–3 The comparative clinical performance of these BBs has not yet been studied. The purpose of this study was to compare the efficacy of these BBs with DLTs for achieving lung collapse. It is generally accepted that BBs perform better clinically when positioned in the left mainstem bronchus (LMB) versus the right bronchus.4 Thus, this study was limited to left-sided thoracic procedures to optimize the conditions for BB function. We further evaluated the time required to achieve correct lung isolation and the number of device repositionings needed to maintain OLV.

Back to Top | Article Outline


After approval by the hospital’s Research Ethics Board, written informed consent was obtained from patients undergoing left-sided thoracotomies or video-assisted thoracoscopic surgery for which periods of OLV exceeding 30 min were deemed necessary. Subjects were consecutive consenting patients scheduled for lung or esophageal surgery with no contraindication to the use of a DLT, BB, or the anesthetic protocol. The patients were randomly assigned immediately before induction of anesthesia to one of the four study groups to undergo OLV with an Arndt® BB (Cook® Critical Care, Bloomington, IN), a Cohen® BB (Cook Critical Care), a Fuji® BB (Fuji Systems, Tokyo), or a left-sided DLT (Mallinckrodt Medical, Cornamadde, Athlone, Westmeath, Ireland). Each of the four lung isolation groups were further subdivided into two subgroups to receive either immediate suction (I) (−20 cm H2O applied at the start of lung isolation) to the DLT bronchial lumen or the BB suction channel or delayed suction (D) (applied 20 min after pleural opening). For the Arndt BB the wire loop was removed to use the channel for suction.

Back to Top | Article Outline

Anesthesia Protocol

After placement of standard monitors including a radial arterial catheter and administration of oxygen (Fio2 = 1.0 for 3 min), anesthesia was induced with propofol (1–2 mg/kg), and maintained with fentanyl (5–10 μg/kg) and one minimum alveolar concentration sevoflurane in oxygen (Fio2 1). Rocuronium 1 mg/kg was used for muscle relaxation with repeat 10 mg boluses as required. Most of the patients had thoracic epidurals placed before induction for postoperative analgesia.

Back to Top | Article Outline

Lung Isolation Methods

The BBs and DLTs were placed by thoracic anesthesia fellows (physicians doing 1 yr of specialized postgraduate training) supervised by staff thoracic anesthesiologists. After induction of anesthesia, the patients’ tracheas were intubated using a Macintosh no. 3 laryngoscope with either a 8.0-mm internal diameter endotracheal tube (ETT) (BB groups) or an appropriately sized left DLT (females <160 cm height: 35F; females >160 cm: 37F; males <170 cm: 39F; males >170 cm: 41F). All BBs were placed via the ETT according to the manufacturers’ recommended techniques using a pediatric fiberoptic bronchoscope (FOB). The BBs were positioned distally in the LMB. The DLTs were passed to an appropriate depth according to the patient’s height5 and tube position was adjusted using a FOB, initially via the tracheal lumen and then via the bronchial lumen.6,7 The tracheal cuff was then inflated. After placement, the cuff of the BB or the bronchial cuff of the DLT was inflated with FOB surveillance and OLV in all patients in the supine position was begun using volume-controlled ventilation (tidal volume of 5–6 mL/kg ideal body weight with the respiratory rate adjusted to maintain the Petco2 in the normal range, Fio2 1). The surgeons were absent from the operating room during DLT or BB placement and were blinded to the airway device by means of a sheet placed over the lung isolation device. The patients were then turned to the right lateral decubitus position for surgery and the position of the lung isolation device confirmed by FOB. The lung isolation device was repositioned if the BB or the bronchial cuff of the DLT was not appropriately positioned in the left bronchus.

Back to Top | Article Outline

Study End Points

The time from beginning of laryngoscopy to lung isolation was recorded. The thoracic surgeons using a verbal analog scale, assessed the lung collapse (lung collapse scores [LCS], 0 = no collapse, to 10 = complete collapse). The LCS was assessed and recorded as they opened the pleura during the procedure (LCS 0), and at 10 min, (LCS 10), and 20 min (LCS 20) after opening the pleura. If the lung collapse was not satisfactory, the FOB was passed to diagnose and correct the problem. The number of repositions of the lung isolation device after initial placement, airway pressures, tidal volumes, duration of surgery, and the arterial blood gases 20 min after pleural opening were recorded.

Other variables that were recorded included the Mallampati score, preoperative pulmonary spirometry results, and the angle of the LMB to the trachea and length of LMB on preoperative chest imaging. The angle and length of the LMB were measured by the anesthesiologist using the on-screen program of the radiology server.

Back to Top | Article Outline

Power Calculation and Data Analysis

Based on the data of Campos and Kernstein8 showing a mean time for placement of the Arndt BB of 200 s, and using a clinically relevant difference in mean placement of 60 s, 25 patients were required per group to have an estimated α error of 0.05 and β error of 0.2.9 To allow for an even number of patients in each suction subsection, 26 patients were recruited to each lung isolation technique group.

Data analyses of LCSs among the groups were with ANOVA (Tukey test was for multiple comparisons). We used analysis of covariance for covariate adjustment with regard to LCS. The number of DLT or BB repositionings was compared using the Fisher’s exact test. Other continuous data were evaluated using the student t-test, and other dichotomous data with Fisher’s exact test. P values <0.05 indicate statistical significance.

Back to Top | Article Outline


One hundred six patients were recruited for the study. Two patients, one in DLT-D group (lung adhesions preventing collapse) and one in the Arndt-I group (inoperable lung cancer) were excluded from the study. These patients were excluded before the study period. Other patients were subsequently recruited and randomized to replace these two subjects. Data were analyzed from 104 patients.

Patient demographics and other preoperative data are listed in Table 1. The groups were similar with regard to age, sex, weight, body mass index, pulmonary function tests, airway grade, LMB angle to the trachea, and types of surgical procedures (Table 2).

Times to lung isolation are shown in Figure 2. The time (mean ± sd) to lung isolation was significantly less for DLTs (93 ± 62 s) than BBs (203 ± 132 s), P = 0.0001. There was no difference between the BBs in the time to lung isolation. The mean time to lung isolation for the Arndt BB was 191 ± 131 s, for the Cohen 204 ± 144 s and for the Fuji 213 ± 144 s (P = 0.78). Lung isolation scores data are shown in Figure 3. There was no difference among four lung isolation devices for LCS at 0 min (P = 0.66), 10 min (P = 0.78), and 20 min (P = 0.51) after pleural opening.

Data on the number of lung isolation device repositionings are shown in Figure 4. There were significantly more frequent repositionings with the BBs (35 incidents) compared with DLTs (two incidents) (P = 0.009). The Arndt BBs needed to be repositioned 16 times, the Fuji BB 11 times and the Cohen BB 8 times. The Arndt BBs needed to be repositioned more often than the other BBs (P = 0.032). Early Suction applied to the Cohen BB (P = 0.018) and DLT (P = 0.019) significantly increased the LCS when compared with late suction but not with the Arndt BB or Fuji BBs (Fig. 5).

There was a significant difference in the mean peak airway pressures between BB (19 cm H2O) and DLT (16 H2O). There was no difference in the mean tidal volumes among the groups (P = 0.66). The pH was significantly lower (P = 0.007) and Paco2 significantly (P = 0.015) higher in the BB group than in the DLT group (Table 1).

Back to Top | Article Outline


Left-sided DLT tubes are commonly used for lung isolation. There has been a recent increase in interest in the use of BBs for OLV after the introduction of new blocker designs and because of the widening scope of surgical procedures requiring lung isolation, such as minimally invasive cardiac procedures and transthoracic spine surgery.10 Although, widely used clinically, DLTs are associated with several limitations, including difficulty with insertion and positioning in patients with abnormal upper or lower airway anatomy and in patients with difficult airways. Further, the use of DLTs requires the change to a single-lumen ETT at the completion of surgery if the patient requires postoperative ventilatory support. This may present challenges in the presence of facial edema, secretions, laryngeal trauma from the initial intubation and large fluid shifts during surgery. Moreover, DLTs, when compared with a single-lumen ETT, are associated with more airway trauma, leading to postoperative hoarseness and throat pain.11 The new BBs, therefore, offer a reasonable alternative to DLTs in such clinical situations. Although the efficacy of right- or left-sided DLT and Univent tubes® has been reported,12,13 there has been no prospective study comparing the effectiveness of lung isolation with the new BBs and DLTs.

This study demonstrated that the tested BBs required an average of 110 s longer to position than DLTs (P = 0.0001) (Fig. 2). However, once lung isolation was achieved, the overall clinical performance, as determined by an objective assessment of lung collapse, was equivalent for all four devices. Further, we found no significant differences among the BBs in time to lung isolation or lung isolation scores. There were more repositionings of the lung isolation device with the BBs compared with the DLTs (P = 0.009). The Arndt BB needed to be repositioned more often that the other BBs (P = 0.032). This may have been due to the elliptical shape of the bronchial cuff in the Arndt BB compared with the spherical cuffs in the Cohen BB and Fuji BB. Subsequent to the introduction of the elliptical Arndt blocker used in this study, introduced a version of the Arndt blocker with a spherical balloon. The spherical blocker is produced in a range of sizes 9, 7, and 5F, whereas the elliptical blocker comes in the 9F size only. There are no published clinical studies comparing the two versions of the Arndt blocker.

The suction channel of the Cohen BB has multiple distal orifices, which may be why early suction improves lung collapse in these patients compared with the other BBs (P = 0.018) (Fig. 5). The larger single bronchial lumen of the DLT (6–9 mm) may explain why early suction showed some benefit (P = 0.019) (Fig. 5). The failure of suction to improve LCS in the Arndt BB in this study is not consistent with results published by Campos and Kernstein.8 A potential criticism of this study is that the level of suction used to try to increase the rate of lung collapse is low. Some clinicians use high levels of suction for this purpose. However, direct wall suction pressures (−200 mm Hg) have been shown to be harmful to airway mucosa.14 We chose to use a suction of −20 cm H2O.

The LCSs were assessed in this study at pleural opening and for the next 20 min, Because we consider this period to be clinically the most relevant. The equivalent rate of lung collapse (Fig. 3) with BBs and DLTs in this study is in disagreement with previous studies.9,15 The improved performance of BBs in this study compared with the other studies of BBs,9,15 in which the Arndt ™ BB took longer to deflate the lung compared with DLT, may be related to the anesthetic protocol. The operative lung was collapsed during a period of apnea and after careful administration of oxygen and ventilation with a Fio2 of 1.0. Residual nitrogen in the operative lung from prior ventilation with air/oxygen mixtures may delay collapse. Finally, we found that the mean airway pressures with the BBs were higher than with the DLTs. Because the tidal volumes were similar, this may indicate that the airflow resistance of a single-lumen ETT with a BB is higher than the tracheal lumen of a DLT.

A limitation of this study is that postgraduate anesthesia fellows performed the lung isolation procedures. This group of trainees was relatively homogeneous in terms of clinical experience. The fellows had more experience with DLTs than BBs before the study and this may be, in part, responsible for the longer time required for lung isolation with the BBs. Experts in thoracic anesthesia have performed most of the clinical studies of BBs.9,11 However, in the study done by Campos and Kernstein8 the time to place an Arndt BB was approximately 200 s, which is comparable with the time it took to place the BBs in this study. The staff thoracic anesthesiologists each had more than 10 yr of clinical experience and were familiar with all devices used before the study. They confirmed the final position of the lung isolation device; therefore, the lack of experience of the operators cannot explain the subsequent increased need for repositioning of the BBs.

In summary, the three BBs provided equivalent surgical exposure to left-sided DLTs during left-sided open or video-assisted thoracoscopic surgery thoracic procedures. BBs required longer to position and required more frequent intraoperative repositioning. The Arndt BB had to be repositioned more often than the other BBs.

Back to Top | Article Outline


1. Cohen E. The Cohen flexitip endobronchial blocker: an alternative to a double lumen tube. Anesth Analg 2005;101:1877–9
2. Arndt GA, DeLessio ST, Kranner PW, Orzepowski W, Ceranski B, Valtysson B. One-lung ventilation when intubation is difficult-presentation of a new endobronchial blocker. Acta Anaesthesiol Scand 1999;43:356–8
3. Nishiumi N, Nakagawa T, Masuda R, Iwasaki M, Inokuchi S Inoue H. Endobronchial bleeding associated with blunt chest trauma treated by bronchial occlusion with a Univent. Ann Thorac Surg 2008;85:245–50
4. Bauer C, Winter C, Hentz JG, Ducrocq X, Steib A, Dupeyron JP. Bronchial blocker compared to double lumen tube for one lung ventilation. Acta Anaesthesiol Scand 2001;45:250–4
5. Brodsky J, Benumof JL, Ehrenwerth J, Ozaki GT. Depth of placement of left double-lumen endobronchial tubes. Anesth Analg 1991;73:570–2
6. Slinger PD. Fibre-optic bronchoscopic positioning of double-lumen tubes. J Cardiothorac Anesth 1989;3:486–96
7. Benumof JL, Partridge BL, Salvatierra C, Keating J. Margin of safety in positioning modern double-lumen endotracheal tube. Anesthesiology 1987;67:729–38
8. Campos J, Kernstein K. A comparison of the left-sided Broncho-Cath with the torque control blocker Univent and the wire-guided blocker. Anesth Analg 2003;96: 283–9
9. Lerman J. Study design in clinical research: sample size estimation and power analysis. Can J Anaesth 1996;43:184–91
10. Campos JH. Progress in lung separation. Thorac Surg Clin 2005;15:71–83
11. Knoll H, Ziegeler S, Schreiber JU, Buchinger H, Bialas P, Semyonov K, Graeter T, Mencke T. Airway injuries after one-lung ventilation: a comparison between double-lumen tube and endobronchial blocker: a randomized, prospective, controlled trial. Anesthesiology 2006;105:471–7
12. Campos JH, Reasoner DK, Moyers JR. Comparison of a modified double-lumen endotracheal tube with a single-lumen tube with enclosed bronchial blocker. Anesth Analg 1996;83:1268–72
13. Campos JH, Massa CF. Is there a better right-sided tube for one-lung ventilation? A comparison of the right-sided double-lumen tube with the single-lumen tube with right-sided enclosed bronchial blocker. Anesth Analg 1998;86:696–700
14. Czarnik RE, Stone KS, Everhart CC Jr, Preusser BA. Differential effects of continuous versus intermittent suction on tracheal tissue. Heart Lung 1991;20:144–51
15. Campos JH. An update on bronchial blockers during lung separation techniques in adults. Anesth Analg 2003;97:1266–74
© 2009 International Anesthesia Research Society