For a patient with a difficult upper airway who requires 1-lung ventilation, one suggested strategy is to intubate the trachea with a single-lumen endotracheal tube (ETT), either awake, initially, or after induction of general anesthesia, and then to exchange the single-lumen tube for a double-lumen endotracheal tube (DLT) over an airway exchange catheter (AEC).1 In our institution, this is typically facilitated with video laryngoscopy to visualize the DLT passing through the glottis.2 However, the bronchial lumen can become impacted at the arytenoid cartilages, making this tube exchange difficult. During fiberoptic intubation, a single-lumen tube may become engaged at the glottis, most commonly becoming impacted on the right arytenoid or the adjacent corniculate cartilage.3 DLTs manufactured by Rusch (Bronchopart®, Teleflex, Research Triangle Park, NC) and Mallinckrodt (Broncho-Cath®, Covidien, Mansfield, MA) have, respectively, a bronchial end that is cut perpendicular to the axis of the tube (i.e., a square cut) or with a very short bevel. The Fuji-Phycon DLT (Broncho-Cath®, Covidien, Mansfield, MA) has a recently modified design with a 45° bevel at the distal bronchial end and a more flexible bronchial tip (Fig. 1). It has been our clinical impression that this design improves the passage of a DLT past the arytenoid cartilages over an AEC, thereby facilitating tube exchange. To test this, we performed a randomized crossover comparison of 3 types of DLTs in a DLT exchange over an AEC with video laryngoscopy using a high fidelity airway simulator.
This was a single-center, randomized crossover trial. It was approved by the University of Toronto, Department of Anesthesia, Postgraduate Education Committee. Eligible participants were anesthesia residents or fellows with at least 3 years of anesthesia training who volunteered to participate in the study and who signed informed consent.
After verbal explanation of the study protocol, each participant performed an oral intubation on a high-fidelity simulator (Airway Management Trainer, model AA-3100, Laerdal, Toronto) using a standard single-lumen 8.0 ETT (Broncho-Cath®, Covidien, Mansfield, MA) and a video laryngoscope (GlideScope, Verathon, Bothell, WA). They then placed a lubricated 11 F/100 cm soft-tipped extra-firm DLT exchange catheter (Cook Inc., Bloomington, IN) through the single-lumen ETT and then removed the ETT. Next, in random order, they passed a size 37F DLT from each manufacturer: Rusch (Bronchopart®, Teleflex, Research Triangle Park, NC); Mallinckrodt (Broncho-Cath®, Covidien, Mansfield, MA); and Fuji-Phycon (Silbroncho®, Fuji Systems, Tokyo, Japan) over the AEC until the bronchial cuff was past the vocal cords, while the glottis was kept in view of the video laryngoscope by one of the authors (RG) holding the video laryngoscope. Randomization of the order of DLT exchange was achieved by blindly pulling the name of the manufacturer out of a box.
The participants were allowed to use any maneuver they would normally use to navigate the DLT into the trachea. Video recordings of each exchange were made from the video laryngoscope by using a Pinnacle Video Transfer System (Corel Corp., Ottawa, ON) and stored on a laptop computer. Intubation times were documented from the recorded video clips at the end of the study. The primary end point was the time to intubate, defined as the time from when the bronchial lumen of the DLT first entered the video laryngoscope view over the AEC, to when the bronchial lumen passed through the vocal cords.
Participants’ ranking of the ease of use of each DLT exchange was documented on a scale from 1 to 10, with 1 being very difficult, 10 being very easy. The failure rate for each DLT was recorded, defined as an attempt >150-second duration, which has been used in previous studies.1 The cause of failure was also documented. Demographic data were also collected from participants, including their experience with the video laryngoscope, fiberoptic intubations, airway exchange catheters, double-lumen endotracheal intubation, and DLT exchange over an AEC.
Sample Size Calculation
This study was powered to show a difference of 20-second intubation times. Based on a pilot study of DLT exchange times by the authors according to the above protocol, with a mean exchange time of 34 seconds, and a standard deviation (SD) of 20 seconds, 17 participants would need to be enrolled to achieve an α error of 0.05 and a β error of 0.2.
Given the nonnormal distribution of the primary outcome (time to intubation) and the small sample size, nonparametric statistical methods were used. Overall effect of DLT period (i.e., order of DLT) and the interaction between DLT and period were determined by calculating and comparing rank statistics within period and pooling over the entire sample. This study did not meet the assumptions of a Latin-squares design for crossover studies (i.e., equal observations for each potential DLT sequence). The methods described by Senn4 were therefore used, and rank sum statistics were pooled across pairwise comparisons for the nonnormally distributed factors. A Bonferroni correction was applied to account for multiple comparisons. To assess differences in whether the intubation was successful (i.e., procedure took <150 seconds) across DLTs, exact permutation tests were used.
The baseline characteristics of all 17 participants screened for eligibility are shown in Table 1. Data for the primary end point as well as the ease-of-use scores were nonnormally distributed and reported as median values (Table 2).
Time to intubate was significantly shorter with the Fuji-Phycon DLT (median 2 seconds) compared with both Rusch (27 seconds) and Mallinckrodt (21 seconds). Overall analysis did not demonstrate a significant period by DLT interaction effect (P = 0.1934) or period effect (P = 0.1387); however, there was a significant difference across DLTs (P < 0.0001). Bonferroni corrected pairwise comparisons of the DLTs indicated that there was a significant increase in median time to intubation for Rusch compared with Fuji (P = 0.0144) and Mallinckrodt compared with Fuji (P = 0.0117). There was not sufficient evidence to demonstrate a significant difference in time to intubation between Rusch and Mallinckrodt DLTs (P = 0.2457).
Failure of Intubation
Exact permutation tests were used to compare DLTs with respect to success or failure of intubation. Rusch DLT was associated with a statistically significant increase in likelihood of failure compared with both Fuji and Mallinckrodt DLTs (P = 0.002 for both comparisons). Specifically, there were 0 (0.0%) failures in simulations using both the Fuji and Mallinckrodt DLTs, compared with 6 (35.3%) failures for the Rusch DLT. In 5 cases, failure was due to the bronchial lumen becoming caught on the right arytenoid, and in 1 case, the airway exchange catheter came out.
Ease of Use
Overall analysis did not demonstrate a significant period by DLT interaction effect (P = 0.0816) or period effect (P = 0.0585); however, there was a significant difference across DLTs (P < 0.0001) with respect to ease of use. Bonferroni corrected pairwise comparisons of the DLTs indicated that ease-of-use scores were significantly lower for Rusch compared with Fuji (P = 0.0186) and Mallinckrodt compared with Fuji (P = 0.0123). There was not sufficient evidence to demonstrate a significant difference in ease of use between Rusch and Mallinckrodt DLTs (P = 0.2364).
This study showed that, in a simulator setting, the Fuji-Phycon DLT was easier to navigate past the vocal cords over an AEC than the Rusch or Mallinckrodt designs. This was due to the distal bevel and flexibility of the bronchial lumen of the Fuji-Phycon model. In reviewing the video images, the variability in intubation times was almost entirely due to impaction of the bronchial lumen against the right arytenoid, rather than becoming caught on the vocal cords or anterior larynx (Figs. 2 and 3). The use of a video laryngoscope allows the operator to orient the distal bevel of the DLT to facilitate passage through the vocal cords. A lack of direct vision at the moment an ETT passes through the glottis is probably one reason that intubation with beveled single-lumen ETTs is occasionally awkward during intubation over a fiberoptic bronchoscope.3,5
With the development of modern disposable polyvinyl chloride DLTs and the increasing use of fiberoptic bronchoscopy to position DLTs, it became evident that the distal beveled bronchial orifice could become partially obstructed by the medial wall of the left mainstem bronchus.6 This led some manufacturers to remove the bevel from the bronchial lumen or to increase the angle between the tracheal and bronchial lumens. The Fuji DLT is made of silicon which is less rigid and has a wire spiral imbedded in the distal bronchial lumen. This increased flexibility seems to allow the distal bronchial lumen to sit more centrally in the mainstem bronchus with less tendency for the orifice to obstruct against the medial wall of the bronchus. Whether this increased flexibility of the distal tip will decrease the risk of bronchial lacerations remains to be determined.
There are several different methods of achieving lung isolation in a patient with a difficult upper airway requiring thoracic surgery. These include the use of bronchial blockers and single-lumen endobronchial tubes. However, DLTs remain the primary choice for achieving lung isolation by most anesthesiologists. In a survey in the United Kingdom, 98% of respondents reported that a DLT was their first choice for lung isolation, and the majority only rarely used bronchial blockers.7 Only DLTs give continuous access to both lungs during lung separation. This is extremely useful for suctioning in cases such as empyema and permits visual confirmation of correct lobar occlusion before bronchial stapling during minimally invasive pulmonary resections. Also, in some clinical scenarios such as endobronchial tumors and bronchial sleeve resections, DLTs are the only practical airway device.8
DLTs can be placed directly either with direct or video laryngoscopy. The usefulness of video laryngoscopes for endotracheal intubation with single-lumen tubes in patients with difficult upper airways is well established. However, it has been the clinical experience of ourselves and others that, although visualizing the glottis with a video laryngoscope is usually straightforward, direct DLT intubation with a video laryngoscope is often awkward in a patient with a difficult airway.2 A specific intubation stylet designed for DLT placement with a GlideScope, the GlideRite DLT Stylet® (CAREstream Medical Ltd., Langley, BC) has been developed to facilitate intubation. This stylet has been reported to be useful in patients with normal airways.9 It has been our experience that this stylet is not always useful in patients with difficult airways.
The use of dual exchange catheters has been described to facilitate passage through the glottis of a DLT.1 One AEC is passed in through the bronchial lumen of the DLT and the other through the tracheal lumen. It requires an ETT of 7.5 mm ID or larger to easily accommodate two 11F AECs simultaneously. An option is to place 2 AECs with smaller ETTs, that is, to place 1 AEC via a small ETT, then remove the ETT, and pass another small ETT beside the first AEC to place the second AEC. However, this is an awkward and time-consuming approach.
For these reasons, our current primary clinical plan in patients with a difficult airway requiring a DLT is intubation with a single-lumen ETT, then a tube exchange to a DLT using an AEC aided by visualization with a video laryngoscope. We use the same procedure with the steps reversed to change back to an ETT at the end of the case if an ETT is required for postoperative ventilation. We have used the Fuji-Phycon DLT for this purpose in >30 patients in the past year, and the clinical performance is similar to the results of this simulator study.
There are several limitations to this study, including those inherent in all simulator studies. It is possible that problems in passing the DLT over the AEC might arise at any point in the airway more proximal than the arytenoids, and our definition of time to intubate would not capture this. However, in our small sample, there were no difficulties in passing the DLT over the proximal AEC, past the tongue or into the pharynx. Similarly, we did not measure time to complete intubation with the DLT because it has been our clinical impression that once the bronchial lumen passes through the glottis, the intubation proceeds without complication if a correct size DLT has been chosen.
Only size 37 DLTs were studied; however, larger tubes would be even more likely to get caught on the arytenoids.
The simulator used in this study was not a specifically designed difficult airway simulator. It is possible that the differences in performance between these DLTs may be more or less in a difficult airway simulator or in a patient with a difficult airway. The technique of video-assisted tube exchange used in this study has been developed for difficult airway management. However, in patients (or simulators) with normal airways, direct laryngoscopy to intubate with a DLT is more efficient.
We did not look at correct placement of the bronchial lumen (i.e., right or left lung), because in a difficult airway situation securing the airway is of utmost importance, and placement corrections can be performed in a less time-sensitive manner. With any ETT-to-left DLT tube exchange using an AEC, our clinical impression is that there is an increased incidence of initial accidental right mainstem bronchial intubation with the DLT. The reasons for this are not certain but could involve a tendency of the AEC to straighten the inherent left lateral curve of bronchial lumen of the DLT and direct it toward the right mainstem bronchus. This accidental malpositioning seems to be independent of the design of left-DLT used. This misplacement can be readily corrected with subsequent fiberoptic bronchoscopy in the large majority of cases. And as with any AEC, care must be taken not to pass the distal tip of the AEC beyond the carina to avoid possible bronchial injury. For an average size adult, this means not passing the AEC deeper than 25 cm at the teeth.
Ideally, a similar comparison of DLT exchanges could be performed on patients. In this study, the longer intubation attempts showed significantly more contact with the arytenoids, which may translate into more trauma in patients. It is difficult to justify repeating this study on patients when this simulator study showed the Fuji-Phycon was clearly quicker to place and appeared to cause less trauma.
All 3 DLTs used in this study are commercially available in North America and were obtained from hospital supplies. One criticism of this study is the cost of the DLTs. The cost per unit DLT of the Rusch and Mallinckrodt will depend on the supplier and contract but can be estimated in the range of U.S. $60 to $75. At present, the cost for the Fuji-Phycon tube is estimated at $110.
A further criticism of this study is the relative inexperience of the study subjects with the technique of DLT exchange over a catheter. Only 5 of 17 subjects had previously performed this procedure on ≥5 occasions in clinical practice. Experienced cardiothoracic anesthesiologists specializing in this area have the clinical opportunity to develop a spectrum of strategies for managing patients with difficult airways requiring lung isolation. However, we believe that the study subjects may reflect the more common clinical practice where an individual anesthesiologist may only infrequently be required to provide lung isolation for a patient with a difficult airway. In this clinical situation, the practitioner would prefer to have the simplest technique and most reliable airway equipment available.
The technique used in this study involved 2 operators (i.e., 2 pairs of hands). The subject first performed the initial ETT video intubation, then handed the video laryngoscope to an assistant, and performed the tube exchange, while the assistant (one of the authors, RG) held the video laryngoscope to show the subject the glottis during the tube exchange. This reflects our clinical practice during which any lung isolation procedure in a patient with a difficult airway, whether with a DLT or a bronchial blocker, requires 2 operators. However, the assistant does not have to be another anesthesiologist. Any member of the operating room team can easily and quickly be taught to hold the view with the video laryngoscope during the tube exchange.
This is the first study to specifically compare DLTs and AECs. The clinical scenario requiring this technique presents infrequently; however, in these cases, one should consider using a Fuji-Phycon DLT.
Name: Ryan Gamez, MD, FRCPC.
Contribution: This author helped design and conduct the study, analyze the data, and write the manuscript
Attestation: Ryan Gamez has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Conflicts of Interest: The author has no conflicts of interest to declare.
Name: Peter Slinger, MD, FRCPC.
Contribution: This author helped design the study, analyze the data, and write the manuscript.
Attestation: Peter Slinger has seen the original study data, reviewed the analysis of the data, approved the final manuscript, and is the author responsible for archiving the study files
Conflicts of Interest: Peter Slinger was an unpaid consultant to Fuji Sytems for the redesign of the of the Silbroncho® double-lumen tube. This author did not participate in data collection or subject recruitment. Neither of the Authors has a current or past financial disclosure relating to any of the devices or manufacturers mentioned in this study.
This manuscript was handled by: Steven L. Shafer, MD.
We would like to acknowledge J. Charles Victor for consulting with us on the statistical analyses.