A tracheogastric tube fistula is a rare complication (0.8%–1.5%) after esophagectomy, but it is frequently life threatening because of respiratory failure, sepsis, or mediastinitis.1,2 Once a tracheogastric tube fistula has been diagnosed, early intervention is required to prevent aspiration. Intervention consists of conservative therapy (e.g., administration of antibiotics, proton pump inhibitors, and total parenteral nutrition) or endoscopic or surgical repair. Among these therapies, a 2-stage surgical repair has been reported as one of the most difficult repairs.3,4
Anesthetic and respiratory management of a patient with a fistula is challenging because air leakage through the fistula and positive pressure ventilation may increase the risk of pulmonary aspiration. Intubation and positive pressure ventilation are sometimes nearly impossible if the fistula is large and positioned at the carina. We report the anesthetic and respiratory management of a patient undergoing 2-stage surgical repair combined with endoscopic mucosal resection (EMR) in a case with a giant tracheogastric tube fistula at the carina.
A 66-year-old woman was scheduled to undergo a 2-stage surgical repair combined with EMR for a giant tracheogastric tube fistula at the carina. Five years prior, she underwent total esophagectomy, lymphadenectomy, thyroidectomy, and pharyngolaryngectomy for cervical esophageal cancer followed by gastric tube reconstruction through the posterior mediastinal route. Although her postoperative course had been uneventful, 5 years after the esophagectomy, she suddenly developed aspiration pneumonia because of a benign tracheogastric tube fistula. Despite administration of antibiotics, proton pump inhibitors, and total parenteral nutrition, insertion of a decompression tube, endoscopic treatment, and direct suturing, the fistula grew until it occupied half of the posterior wall of the trachea and the carina collapsed (Fig. 1). A 2-stage surgical repair combined with EMR was indicated to separate the food passage from the trachea and to reconstruct the digestive tract.
The patient was 143.1 cm tall, weighed 38.9 kg, and was ASA physical status II. Laboratory findings and arterial blood gas were within normal limits with the exception of a C-reactive protein level of 2.37 mg/dL.
The first surgical stage was division of the gastric tube above the fistula and establishment of an external gastric tube and salivary fistulas using a cervical incision (Fig. 2). The estimated operative duration was 1 hour. The anesthesia plan was local anesthesia and light sedation with minimal opioids to maintain spontaneous breathing guided by bispectral index monitoring. We did not choose general anesthesia with 1-lung ventilation because of the risk of destruction around the fistula by a tracheal tube.
After the patient entered the operating room, 35 mg meperidine and 2.5 mg diazepam were administrated IV to complete a final bronchoscopic examination of the position and size of the fistula. A wire-reinforced tube (Mallinckrodt™ oral/nasal tracheal tube cuffed, 6.0-mm ID; Covidien Japan, Tokyo, Japan) without its cuff inflated was placed in the tracheostomy orifice above the fistula before beginning surgery. One-lung ventilation with this tube was planned as an emergency respiratory backup in case spontaneous respiration ceased. The patient was sedated with an infusion of 2 to 4 mg/kg/h IV propofol and 3 mL 1% lidocaine was injected into the operative field. Upon incision, the bispectral index was 70, spontaneous breathing was maintained, and no body motion was observed. When the patient’s eyes opened, an infusion of fentanyl 0.04 mg/h was started. An IV bolus of fentanyl 0.01 mg was administered twice, and an additional 1% lidocaine was injected into the operative field as needed. Suctioning of blood and secretions from the decompression tube were performed several times to prevent aspiration. The spontaneous respiratory and hemodynamic variables remained stable during surgery, which lasted 78 minutes. After the first surgical stage, no aspiration pneumonia was observed.
Before the second surgical stage, multiple EMRs were attempted to obstruct the gastric tube below the fistula to block gastric acid reflux because healing of iatrogenic ulceration after EMR causes a cicatricial stricture to form; however, a pin hole that could be detected by endoscopy remained. To reduce acid production, lansoprazole 15 mg daily was given via an enteral tube that had been inserted at the time of the direct suturing surgery. The second surgical stage consisted of closure of the gastric tube above the fistula, insertion of a decompression tube in the gastric tube below the pin hole, and reconstruction of the digestive tract using the jejunum (Fig. 3). The anesthetic and respiratory management plan was to preserve spontaneous breathing with pressure support under general anesthesia and to achieve abdominal muscle relaxation using an epidural, avoiding use of a muscle relaxant to avoid high positive pressure until the abdominal cavity was opened and a decompression tube inserted.
After the patient entered the operating room, an epidural catheter was inserted at the ninth and tenth thoracic interspace. After administering 6 L/min of oxygen via the tracheostomy orifice, 35 mg of meperidine, and 5 mg of diazepam IV, a wire-reinforced tube (Mallinckrodt™ oral/nasal tracheal tube cuffed, 8.0-mm ID; Covidien Japan) was inserted into the trachea above the fistula under spontaneous ventilation. Anesthesia was maintained with an infusion of 2 to 5 mg/kg/h propofol and 0.05 mg fentanyl. Spontaneous ventilation was facilitated with 5 cm H2O of pressure support. Before the abdominal incision was made, 7 mL of 0.3% ropivacaine was injected epidurally followed by continuous infusion at a rate of 6 mL/h. The orifice of the external gastric tube fistula was closed. Gastric distention was not observed when the abdominal cavity was opened. After a decompression tube was inserted into the gastric tube below the stenosis to prevent regurgitation of gastric acid, 30 mg rocuronium was administered and intermittent positive pressure ventilation was initiated. The observed respiratory and hemodynamic variables remained stable during surgery. Upon completion of surgery, the tracheostomy was reestablished. After the second surgical stage, no aspiration pneumonia was observed. On day 33, the patient was discharged in good condition, but on day 36 she was rehospitalized because of minor leakage. Gastroscopy revealed a small fistula at the cervical anastomosis of the gastric tube and the jejunum. We closed the fistula under local anesthesia and propofol sedation. The patient was finally discharged in good condition on day 31 after this last surgical procedure.
There are few reports about 2-stage surgical repair of a tracheogastric tube fistula,3,4 and the anesthesia and respiratory management for this surgical procedure have not been well described to our knowledge. Because our patient had a giant fistula at the carina and the operative procedure did not close the fistula itself, the anesthesia and respiratory management needed to be carefully devised. When planning the anesthetic management, it was very important to understand the surgical plan and the patient’s anatomy.
Because surgical stress was low in the first stage, local anesthesia and sedation were considered sufficient. The patient awakened twice but did not complain. General anesthesia may not always be necessary in this stage. Positive pressure ventilation under general anesthesia could be an option for patients with a tracheal fistula with enough distance between the fistula and the carina because a tube cuff can easily block the fistula itself or the airway below the fistula.2 Previous reports have shown that 1-lung ventilation with a small endotracheal tube,5 independent ventilation of each lung with 2 endotracheal tubes,5 or high-frequency jet ventilation6 were useful anesthetic approaches in a patient with a giant fistula at the carina similar to ours. We believe that the combination of local anesthesia and sedation is advantageous because it poses minimal risk of destruction around the fistula by a tracheal tube or high pressure. This is because the 2-stage surgical repair does not close the fistula itself nor can it provide access around the fistula. In addition, our patient had a history of direct suture failure because of disturbed blood circulation. Therefore, the destruction around the fistula could have led to irreparable damage.
Before the second surgical stage, multiple EMRs were attempted to obstruct the gastric tube below the fistula to block gastric acid reflux. This was attempted to induce the formation of cicatricial strictures that often occur after the post-EMR iatrogenic ulceration heals. In these surgeries, it was unavoidable that a diverticular gastric tube would remain in the airway. Therefore, it was very important to make this diverticular tube as small as possible because it represents ventilatory dead space. If a large diverticular gastric tube had been left, aspiration of large amounts of acidic gastric contents might have caused death postoperatively. We considered multiple EMRs as a less invasive approach than open chest surgery to create an obstruction near the fistula. Although conventional intermittent positive pressure ventilation could have been performed under perfect occlusion of the gastric tube below the fistula, our selected respiratory management was spontaneous ventilation with pressure support to maximally reduce positive pressure in the second surgical stage because a pin hole remained.
After opening the abdominal cavity and inserting a decompression tube, we chose to administer rocuronium and begin mechanical ventilation. This enabled us to clamp the abdominal gastric tube at any time if the digestive tract inflated. We believed that the absence of dilation of the digestive tract at the time the abdominal cavity was opened confirmed that there was no communication through the pinhole at the existing pressure gradient. We anticipated that using a decompression tube might reduce the risk of aspiration. Gastric distention was absent during the surgery, and no aspiration was permitted after the surgery under this management plan. Although relaxation of the abdominal muscles was obtained through the use of epidural anesthesia, this method cannot be used for patients in whom epidural anesthesia is contraindicated.
In conclusion, we present a novel approach to the anesthetic management of patients undergoing the above-described surgical procedure for a giant tracheogastric tube fistula. In the first stage, the strategy was local anesthesia and sedation, and in the second stage, pressure support under general anesthesia with epidural anesthesia was effective.
Consent for Publication
Written informed consent was obtained from the patient for publication of this case report.
1. Bartels HE, Stein HJ, Siewert JR. Tracheobronchial lesions following oesophagectomy: prevalence, predisposing factors and outcome. Br J Surg. 1998;85:403–6
2. Yasuda T, Sugimura K, Yamasaki M, Miyata H, Motoori M, Yano M, Shiozaki H, Mori M, Doki Y. Ten cases of gastro-tracheobronchial fistula: a serious complication after esophagectomy and reconstruction using posterior mediastinal gastric tube. Dis Esophagus. 2012;25:687–93
3. Tsujinaka T, Ogawa M, Kido Y, Shiosaki H, Mori T. A giant tracheogastric tube fistula caused by a penetrated peptic ulcer after esophageal replacement. Am J Gastroenterol. 1988;83:862–4
4. Freeman MS, Livingstone AS, Goodwin WJ Jr. Giant acquired tracheoesophageal fistulas: strategy for successful management. Head Neck Surg. 1986;8:463–5
5. Roy JS, Girard F, Boudreault D, Pinard AM, Ferraro P. The anesthetic management of a case of tracheogastric fistula. Anesth Analg. 2001;93:1076–7
6. Tsui SL, Lee TW, Chan AS, Lo JR. High-frequency jet ventilation in the anesthetic management of a patient with tracheoesophageal fistula complicating carcinoma of the esophagus. Anesth Analg. 1991;72:835–8