A COMPLETE tear of the posterior membrane of the trachea due to extubation of a single-lumen (SLT) or double-lumen tube (DLT) has never been described. The facts of the case reported herein show that extubation of a DLT in a patient with persistent complete collapse of the left upper lobe likely caused a 7-cm long, clean, midline tear of the posterior membrane of the trachea.

The patient was a 76-yr-old, 153-cm, 63-kg woman who had the appearance of two new 1-cm opacifications in the left lung and was scheduled for thoracoscopic lung biopsy. The patient was on no medications, was allergic to erythromycin, had a remote history of hysterectomy, and had no significant present medical problems. On physical examination, the patient appeared frail, but all organ symptoms functioned within normal limits, and all laboratory values were within normal limits.

General anesthesia was induced with 2 mg/kg intravenous propofol, 5 μg/kg fentanyl, and 0.05 mg/kg midazolam and was maintained with 0.5–1.0% isoflurane. Paralysis was induced with 0.1 mg/kg vecuronium. After paralysis, a styletted, 8.0-mm inside diameter SLT was easily inserted (the tip of the stylet was 2 cm from the tip of the SLT and was entirely intraluminal) with the aid of conventional laryngoscopy to a depth of 20 cm at the incisors, and fiberoptic bronchoscopy (5.0-mm outside diameter) revealed a normal tracheobronchial tree. The distal 7 cm of the trachea was available for inspection. Next, a 35-French, left-sided, relatively blunt-tipped DLT (Catalog #95895; Mallinkrodt Medical, St. Louis, MO) with stylet was inserted. The stylet was entirely within the left lumen and was removed as soon as the left endobronchial cuff passed the vocal cords and before the 90° counterclockwise rotation of the DLT. Unilateral clamping and auscultation maneuvers revealed normal breath sounds and cannulation of the left lung with the endobronchial lumen of the DLT. Fiberoptic bronchoscopy (4.0-mm outside diameter) through the tracheal lumen revealed that withdrawal of the DLT by 2 cm was required in order to place the blue endobronchial cuff just below the tracheal carina. With the blue endobronchial cuff just below the tracheal carina, the distal 4 cm of the trachea was visualized and was found to be normal. Both tracheal and endobronchial cuffs were inflated so that the pilot balloons were soft (significantly compressible between the fingers with minimal effort), and there was no air leak by either cuff at a pressure of 30 cm H2O, as demonstrated by auscultation.

The patient was disconnected from the ventilator and then easily turned to the right lateral decubitus position without deflation of the softly inflated cuffs and without any movement of the head and neck. For the next 75 min, only the right lung was ventilated with a tidal volume of 500 ml, a respiratory frequency of 12 breaths/min, a peak inspiratory pressure of 30 cm H2O, and a fraction of inspired oxygen of 1.0. During these 75 min, the surgeon biopsied and stapled shut a site in the lingula and left lower lobe. Prior to withdrawing the thoracoscope, an attempt was made to expand the collapsed left lung with a large tidal volume and sustained peak inspiratory pressure of 40 cm H2O for 10 s. However, the left upper lobe (LUL) remained completely atelectatic. The LUL remained collapsed after both conventional suctioning down the left lumen of the DLT and after saline lavage and suctioning of thick, bloody secretions in the LUL through a fiberoptic bronchoscope (4.0-mm outside diameter) in the left lumen of the DLT.

Next, the patient was disconnected from the ventilator and easily turned to the supine position with both pilot balloons to the cuffs of the DLT still softly inflated and without any movement of the head and neck. The patient was then reventilated with manual positive pressure ventilation for 2 min. The cuffs of the DLT were then completely deflated, and the DLT was withdrawn. During extubation of the DLT, no resistance to pulling the DLT out was noted. An 8.0-mm inside diameter, styletted SLT was easily inserted once again to a 20-cm depth at the incisors, and the cuff was inflated so that the pilot balloon was soft and no leak auscultated over three manual breaths. Fiberoptic bronchoscopy (5.0-mm outside diameter) down the new and second SLT (in order to lavage and suction the LUL) very surprisingly revealed a 7-cm long, complete, midline, clean tear of the membrane of trachea that began 2 or 3 mm above the carina and ended 7 cm cephalad. The distal 1 cm of the tear veered slightly to the right. Along the entire length of the tear, the esophagus could be seen to be bulging slightly into the lumen of the trachea. The peak inspiratory pressure was reduced to 12 cm H2O, the respiratory frequency was increased to 20 breaths/min, and the LUL was quickly lavaged and suctioned with a larger fiberoptic bronchoscope.

A right trapdoor thoracotomy with slight left lateral body tilt was performed, the SLT was guided both fiberoptically and from within the chest into the left mainstem bronchus, and the tracheal tear was repaired by primary suture and secondary reinforcement with a pericardial flap. During the repair of the trachea, the tissues of the trachea were felt to be of normal strength and integrity. At the end of the repair of the tracheal tear, the distal end of the SLT was positioned just proximal to the proximal end of the sutured tear, and the patient was sent to the intensive care unit for postoperative mechanical ventilation. Pathologic diagnosis of the biopsy specimens was benign granuloma.

The patient was mechanically ventilated and bronchoscoped daily for 1 month prior to extubation. For the next month after the initial extubation, the patient had repeated bouts of aspiration and required reintubation and mechanical ventilation three times. After the third reintubation, the patient had a tracheostomy. Nevertheless, she died 4 months later, malnourished and septic (perhaps from infection around a jejunostomy feeding tube).

The tracheal tear most likely occurred between the end of the thoracoscopy and the insertion of the second SLT. There are two very important reasons why the tracheal tear could not have been present during and at the end of the thoracoscopy. First, when the DLT was first inserted in the supine position, the distal 4 cm of the trachea appeared normal. Second, it is extremely unlikely that the patient could survive 75 min of positive pressure ventilation with a peak inspiratory pressure of 30 cm H2O in the lateral decubitus position without any air entering the mediastinum; indeed, the surgeon had a good view of the mediastinum, and no abnormality was noted.

The tracheal tear had to be present prior to the insertion of the second SLT. As soon as the second SLT was easily inserted and after approximately three manual positive pressure ventilations at a fraction of inspired oxygen of 1.0, almost all of the tracheal tear was visualized below the tip of the second SLT.

Only two events occurred between the end of the thoracoscopy and the insertion of the SLT—the patient was turned to the supine position, and the DLT was extubated from the trachea. It is very unlikely that turning the patient to the supine position caused the tear, because it was easily and gently accomplished, with no movement of the head and neck, and the pilot balloons of both cuffs were soft. Following the turn to the supine position, no abnormalities in positive pressure ventilation through the DLT were noted for the next 2 min.

Consequently, from a timing point of view, the extubation of the DLT very likely caused the tracheal tear. We hypothesize that the loss of left hemithorax volume due to LUL collapse pulled the posterior membrane of the tracheal tight in the carinal area so that the posterior membrane lost its normal distensibility. During extubation, the tip of the relatively blunt left endobronchial lumen most likely rotated posteriorly enough to penetrate the now taut posterior membrane and slice its way through the membrane until the normal distensibility of the membrane and rotation of the DLT allowed the tip of the left lumen to disengage from the membrane. In support of this hypothesis, a midline incision into a posterior membrane that is pulled to the left would be expected to veer to the right after relaxation of the posterior membrane by the incision (and/or subsequent reexpansion of the LUL). Furthermore, the midline incision appeared sharp and clean as opposed to the cuff-induced tears, which tend to appear ragged, bleed, and occur at the right junction of the membrane and the tracheal cartilages. 1 

Heretofore, the parts of the DLT that have been considered to be culprits for tracheobronchial tree damage have been the endobronchial tip on intubation, a projecting stylet from the endobronchial lumen, and overinflation of either of the DLT cuffs (including diffusion of nitrous oxide into the cuff). 1–3This case is important because it suggests that extubation of a DLT in the presence of a complete lobar collapse may also create a risk for a tracheal tear. This makes sense because it seems logical that the tip of the DLT would have just as much a chance of causing damage in an area of pathology going in one direction as in the other direction. In this case, the pathology (complete LUL collapse causing loss of distensibility of the posterior membrane in the carinal region) was created post-DLT intubation and pre-DLT extubation. Experiments in animals under these specific conditions will be necessary to more precisely define the risks of such pathology.

Marty-Ane CH, Picard E, Jonquet O, Mary H: Membranous tracheal rupture after endotracheal intubation. Ann Thorac Surg 1995; 60: 1367–71
Fitzmaurice BG, Brodsky JB: Airway rupture from double-lumen tubes. J Cardiothorac Vasc Anesth 1999; 13: 322–9
Kaloud H, Smolle-Juettner FR, Prause G, Franz W: Iatrogenic ruptures of the tracheobronchial tree. Chest 1997; 112: 774–8