Background

Anastomotic leakage is a rare but life-threatening complication of esophagectomy for esophageal cancer or esophagogastric carcinoma [1, 2], with an overall mortality rate of 20 to 50% [3,4,5,6]. Contamination in the abscess cavity may corrode vessels and even result in a higher mortality [7]. Various conservative treatment protocols have been used for the management of anastomotic leakage over the past two decades, including the application of biodegradable fistulae plugs or fibrin glue, endoscopic transluminal drainage or clipping and metallic esophageal stent insertion [4, 8,9,10]. Despite these modalities, management of anastomotic leakage remains challenging and the optimal treatment protocol need to be determined [2, 4, 11].

In this study, three-tube method (abscess drainage tube, jejunal feeding tube and gastrointestinal decompression tube) and esophageal covered stent was used. We aimed to determine the safety, feasibility and efficacy of this protocol for the management of anastomotic leakage.

Methods

Patient selection

This study was approved by the Ethics Committee Board of Zhengzhou University First Affiliated Hospital. Informed consent was obtained from all patients. This study enrolled all patients with anastomotic leakage after esophagectomy who received three-tube method and esophageal covered stent placement in our institution between April 2011 and July 2018. The diagnosis of an anastomotic leakage was made based on esophagography (Fig. 1a-b) and chest computed tomographic scan (Fig. 1c–d). During the observation period, there were no changes in technique. Three-tube method was used due to financial difficulties or the position of the fistula was not suitable for stent placement. Patients were divided into 2 groups dependent on the size of the leak: small leaks (less than 6 mm) and large leaks (more than 6 mm).

Fig. 1
figure 1

A 59-year-old man with anastomotic leak after esophagojejunostomy (case 24). a-b Esophagography showing an anastomotic leakage in the lower esophagus and irregular abscess in the pleural cavity. c–d Chest CT scan in the mediastinal and lung windows show mediastinal abscess and a small amount of pleural effusion before procedure

Full size image

Three-tube method

All interventional procedures were performed under fluoroscopic guidance, local anesthesia and conscious sedation. The esophagus and pharyngonasal cavity was anesthetized with an oral lidocaine gel and a tetracaine spray, respectively. A 5-F cobra catheter was introduced through the outlet of anastomotic leakage and into the abscess cavity. The catheter tip was inserted into the distal end of the abscess cavity, followed by exchange with a 5-F pigtail or straight catheter (Cook Medical, Inc., Bloomington, IN). Continuous negative pressure suction was performed using a 20 ml syringe thereafter, with an appropriate pressure to drain the abscess at the same time avoiding excessive negative pressure in order not to cause iatrogenic injury and bleeding. We reduced negative pressure or stopped aspirating in case of bleeding. The abscess cavity was repeatedly rinsed with 100–200 ml of saline. A jejunal feeding tube and gastrointestinal decompression tube was inserted into the jejunum and gastric cavity, respectively. Enteral nutrition solution was infused via the jejunal feeding tube. Patients were allowed to resume oral intake once the leaks had been sealed by the covered stent and confirmed by contrast study. Broad-spectrum antibiotic treatment was performed before and after procedure.

Esophageal covered stent placement

All patients received fluoroscopic placement of esophageal covered stent (Nanjing Micro-Tech Medical Company, Nanjing, China). The stent diameter ranged from 18 to 22 mm and stent length ranges from 70 to 160 mm. A 5-F catheter was inserted transorally into the gastric cavity and then a stiff guide wire was introduced. A covered stent was delivered via the stiff guide wire and then released carefully to block the leakage. A stent fixation line was used for fixation and stent adjustment. Leakage closure was confirmed by repeated esophagography (Fig. 2a-b). About 5 to 7 days after stent placement, chest CT and esophagography were performed again to study whether esophageal stent fit watertight, as well as the change of the abscess cavity and the position of abscess drainage tube (Fig. 2c-d). Stent adjustment or second stent placement were performed if esophageal stent does not fit watertight. The tube was adjusted to make sure effective drainage if necessary. The drainage tubes and stents were removed if complete disappearance of abscess cavity and full expansion of the lungs was confirmed by esophagography (Fig. 3a-b) and chest CT showed (Fig. 3c-d).

Fig. 2
figure 2

Three-tube method and covered stent placement for case 24. a After esophageal covered stent and drainage tube placement, esophagography shows that the contrast agent flows though the esophagus and stent with no leakage. b Esophagogram via drainage tube showing a decreased abscess cavity during follow up. c–d At 1.6 months after three-tube treatment and stent placement, a chest CT scan shows decrease in mediastinal abscess and pleural effusion

Full size image
Fig. 3
figure 3

Examination after stent removal for case 24. a-b The drainage tube and stent were removed after procedure; esophagography shows that the contrast agent flows though the esophagus without any leakage. c–d The chest CT scan shows disappearance of mediastinal abscess, full expansion of the lungs with no pleural effusion

Full size image

Results

General information

This study involved a total of 24 patients with anastomotic leakage, including 21 men and 3 women (Table 1). The age range of patients was 43 years to 76 years, with a median age of 59.5 years. Half of patients showed normal temperature after esophagectomy, the remained patients showed fever, with a median temperature of 38.6 °C (range: 37.5 to 40.0 °C). The median disease course before referral to our department was 3.0 months (range: 0. 3 to 12 months). The median interval between esophageal surgery and leakage was 0.4 months (range: 0.2–9.0 months). There were 16 cases of gastroesophageal anastomotic fistula and 8 patients showed anastomotic leaks after esophagojejunostomy.

Table 1 Patients characteristics of this study
Full size table

Interventional procedure outcomes

Three-tube method was performed successfully for all patients (100%), only 1 patient failed Y stent placement due to complete occlusion (95.8%). A total of 31 esophageal covered stents were placed, with a median diameter of 20 mm (range: 18–22 mm), median length of 120 mm (range: 70–160 mm). Stent placement was successful in the remaining 23 patients, with satisfactory expansion and appropriate position. For those patients, all anastomotic leakages were completely blocked and all stents fit watertight after covered stent placement confirmed by immediate post procedural esophagography. Three patients were in need of external thoracic drainage. All patients showed a reducing amount of drainage fluid, approximately 20 to 250 mL per day. Body temperature returned to normal within 2 weeks after the interventional treatment for patients with fever.

Complications

No perioperative death, esophageal rupture, massive hemorrhage, or other severe complications were observed during procedures. No intensive care was needed due to a worsening condition after the procedure. No necrosis due to negative pressure effects on the esophageal wall was observed after stent placement. Stent migration, the most common complication, was found in 9 patients, with a migration rate of 39.1% (9/23). Three patients showed stent restenosis, with a restenosis rate of 13.0% (3/23). All migrated or restenosed stents were adjusted or replaced for 1 to 3 times (median: 1.0 time). Migration of abscess drainage tube was found in 1 patient. The abscess drainage tube was adjusted or replaced for 0 to 6 times (median: 2 times).

Follow-up

All patients were successfully followed up, with a median time of 7.9 months (range: 1.2–147.6 months). Chest CT and esophagography showed that the abscess cavity had markedly decreased in 8 patients or disappeared in 16 cases. A higher rate of abscess cavity resolution was found in smaller leaks compared to that of large leaks (6/12). However, this different was not statistically significant (P = 0.11), due to small sample size. During follow up, esophageal stents and abscess drainage tubes were successfully removed from 14 patients. The remaining patients refused to remove stent due to heavy financial burden and fear of possible risks of removal. The median retention duration was 2.3 months (range, 0–14.3 months) for stents and 2.6 months (range, 0.4–18.6 months) for abscess drainage tubes, respectively. To this date 14 patients are still alive, with 12 patients returning to their normal daily activities of living and symptom free and 1 patient with slight reflux which is not requiring any treatment. During follow up, 8 patients died of cancer recurrence and 2 patients died of severe pulmonary infection. The 1-, 3-, 5-year survival rates were 60.1, 51.5 and 51.5%, respectively (Fig. 4). A higher rate of 1-year survival rate (70.7%) was found in small leaks than that of large leaks (48.6%). However, this different was not statistically significant (P = 0.31).

Fig. 4
figure 4

Survivals follow up. The 1-, 3-, 5-year survival rates were 60.1, 51.5 and 51.5%, respectively

Full size image

Discussion

Management of anastomotic leakage is challenging for patients received esophagectomy The mortality is high, however, the optimal treatment still need to be determined [2, 4, 11]. Various conservative treatment protocols have been reported for the treatment of anastomotic leakage over the past two decades, including the application of biodegradable fistulae plugs or fibrin glue, endoscopic transluminal drainage or clipping and metallic esophageal stent insertion [4, 8,9,10]. Surgical repair of the anastomotic leakage is the traditional protocol, such as, esophagectomy or thoracotomy and repair [12]. Despite of the advances in surgical technique, the overall mortality remains as high as 20 to 50% [3,4,5,6]. Nowadays, treatment of choice in the first line is an endoscopic approach.

Esophageal stents were initially served as a palliative treatment for patients with esophageal carcinoma. Currently, esophageal stents have been used to treat esophageal benign diseases [13,14,15,16,17]. Metallic stents are usually used for treatment of tumor stenosis or esophageal perforation [18, 19]. Successful and effective management of anastomotic leakage needs prompt elimination of contamination by covered stent placement, and adequate drainage of the abscess cavity. We present 24 consecutive patients treated with three-tube method and covered stent placement for anastomotic leakage. Our clinical outcomes indicated that this interventional method can easily be performed under fluoroscopic guidance. No perioperative death was observed, which is lower than previous reports [6, 15, 16, 20]. After covered stent placement, the leakage is still allowed to continuously drain fluid via abscess drainage tube. In our study, all patients received continue abscess drain for a median duration of 2.6 months. Drainage of an abscess cavity is also possible percutaneously under CT scan control and it is generally easier for the subsequent follow-up. In this study, three patients were in need of external thoracic drainage. Compared with endoscopic drainage, radiologic drainage can be performed via the leaks without the need of percutaneous puncture.

The duration between esophageal surgery and leakage diagnosis is essential for the clinical outcomes [12]. The median interval between surgery and leakage was 0.4 months. The clinical outcomes were favorable; abscess cavity was markedly decreased in 8 patients and disappeared in 16 cases. Treatment results may be related to the size of the leak, and effective results of biodegradable fistulae plugs or fibrin glue application are generally observed only in small leaks.

Certain complications can be found in our interventional protocol. Stent migration is a common complication, especially in patients without esophageal stricture [21, 22]. All esophageal stents used in our study were coved ones, which may account for high rate of stent migration. Nine patients showed stent migration and were adjusted or replaced for 1 to 3 times. Only 1 patient showed migration of drainage tube, however, drainage tubes were regularly adjusted and replaced for 0 to 6 times to achieve effective drainage during follow up. The abscess drainage tubes were adjusted or exchanged for a median time of 2 times. Besides, esophageal stents and abscess drainage tubes were successfully removed from 14 patients, without difficulties of removal or severe complications. Recovery lines in the proximal end of stent are used for stent fixation to avoid stent migration, and for the adjustment or recovery of the migrated stent. Appropriate size of stent should been used, considering that small size of stent is prone to migrate. Of course, clips and stent with flaps can be used to reduce the migration rate.

There were some limitations. This was a retrospective study with relatively small number of enrolled patients. The esophageal stents and abscess drainage tubes needed adjusted or replaced repeatedly during follow up. BMI data had not collected previously considering that BMI may be not closely related to our treatment. We had not measured how much negative pressure and the study interval was long.

Conclusions

Three-tube method and covered stent placement can be considered a safe and effective alternative to operative treatment for anastomotic leakage after esophagectomy. Combined interventional protocol with additional supportive therapy is useful to achieve good clinical outcomes.