We have herein described a patient who developed a fistula between the esophagogastric anastomosis and cartilage portion of the trachea after esophagectomy for cervical esophageal cancer. Fistulae between the esophagogastric anastomotic site and trachea or bronchus after esophagectomy have been reported, but all occurred at the membranous portion of the tracheobronchus. The anastomotic site is often closed to the trachea anatomically, and an esophagotracheal fistula could occur at the fragile membranous portion of the trachea. In the current case, the gastric tube was pulled up through the retrosternal route, and the high cervical anastomosis was adjacent to the left side tracheal cartilage, not to the membranous portion (Fig. 3b). The anastomotic site was obviously higher than the usual anastomosis for thoracic esophageal cancer. We believe that this is the main reason why the esophagotracheal fistula formed at the cartilage portion, not at the membranous portion. This point also represents a significant anatomical difference between the current case and previously reported case involving a fistula at the membranous portion [4–9]. When we perform gastric tube reconstruction through the posterior mediastinal route and anastomosis with a linear stapler, the staple line on the lesser curvature of the gastric tube and anastomotic site can contact the membranous portion of the trachea. Because the membranous portion of the trachea is fragile, we routinely prevent the staple line from contacting the trachea. To accomplish this, the staple line on the gastric tube and anastomotic site is covered with seromuscular sutures and the greater omentum is interposed between the staple line and membranous portion of the trachea. However, when reconstruction is performed through the retrosternal route, the staple line usually does not contact the membranous portion of the trachea. In this case, we only covered the staple line located on the lesser curvature of the gastric tube with seromuscular sutures. The anastomotic site was not covered with seromuscular sutures, and the greater omentum was not interposed because we used a circular stapler; theoretically, the staples were not exposed outside of the anastomotic site.
According to some reports, the incidence of anastomotic leakage after esophagectomy ranges from 10 to 35 % [11, 12]. The degree of intraoperative gastric ischemia due to gastric tube creation is associated with the development of anastomotic complications . Anastomotic leaks are responsible for approximately 40 % of post-esophagectomy deaths . In the current case, we suspected that neoadjuvant chemoradiotherapy and high cervical anastomosis may have been the factors that caused the anastomotic leakage [14–16]. Neoadjuvant chemoradiotherapy is an independent risk factor for anastomotic leakage, as reported in previous literature [14–16]. Neoadjuvant chemoradiotherapy might affect the blood supply, immune system, or fibrosis of tissues, which might result in anastomotic leakage . Peritracheal inflammation due to anastomotic leakage may result in a fistula to the trachea, which is probably the most important course of this complication . Clinically, apparent thoracic anastomotic leaks and fistulae are associated with a high rate of mortality in spite of advances in critical care . In this case, we considered the following possible mechanisms of development of the esophagotracheal fistula. First, inflammation caused by anastomotic leakage spread to the trachea. Second, the staples originating from the circular stapler used for the anastomotic site or from the linear stapler used for the lesser curvature became exposed following the anastomotic leakage and directly injured the trachea. In addition, the trachea was fragile or might have been injured because preoperative chemoradiotherapy might make intraoperative adhesiotomy difficult. When the anastomosis is adjacent to the trachea, even the cartilage portion of the trachea, vital tissue may need to be interposed between the trachea and anastomosis.
Although this is a serious complication, there is no standard management of esophagotracheal fistula. Therapy should depend on the site, size, underlying cause, and severity of symptoms. A conservative treatment regimen may be considered for patients with only mild symptoms. When conservative treatment fails or symptoms are more severe, surgical intervention is necessary. According to the literature, if the fistula fails to heal within a 4- to 6-week period, conservative treatment should be abandoned. Our patient developed a persistent cough and fever. Blood tests revealed prolonged inflammation. The patient’s respiratory condition gradually worsened, and he was considered to have pneumonia, pulmonary sepsis, and respiratory failure. We decided to perform an operation. An omental or pleural patch or a muscle flap can be applied to fill the dead space and add vital tissue to the defect, preventing recurrent fistulization [18, 19]. Because blood flow is poor in the area of the leakage, these vital soft tissues play an important role in wound healing and control of local infection. In our institute, a pectoralis major muscle flap has been used to repair prolonged anastomotic leakage after esophageal reconstruction, and we have been able to manage this serious complication successfully up to the present time . The advantages of the pectoralis major muscle flap for head and neck reconstruction are well known; it is a readily available source of vascularized tissue and is easily harvested for use in the head and neck.