Anastomotic complications after upper gastrointestinal surgery represent an arduous therapeutic challenge. Avoiding a revisional surgery is often desirable and might be possible with an increasing use of stents placed endoscopically [1]. This strategy of management for anastomotic complications is effective in resolving symptoms and expediting enteral nutrition, and turns out particularly successful for sealing anastomotic leaks. In the absence of stents especially dedicated for surgically altered gastrointestinal anatomy, the most appropriate stent selection might reduce the risk of migration.

Fully covered self-expanding metal stent placement is an established treatment for malignant dysphagia and an emerging therapeutic option for benign esophageal strictures. However, the migration is the most frequently reported undesirable event in the literature with covered stents, varying from 3 [2] to 83 % [3]. This high migration rate may be explained by the fact that such stents are not designed specifically for the indication of esophageal leakage or fistula. Moreover, they are completely covered to facilitate their withdrawal and avoid a stenosis [4].

Thus, the use of partially covered stents is a more recent alternative that could significantly decrease the migration rate [5]. However, although they have been used successfully for malignant indications, they have been associated with significant drawbacks in benign indications [6–8]. The most important issue is the hyperplastic tissue reaction of the esophageal mucosa induced by the bare metal mesh. That may result in a recurrent dysphagia due to a stricture at both the uncovered stent edges, and in issues for the removal as well. Nevertheless, Deviere’s et al. and Swinnen et al. in 2 major series [9, 10] demonstrated that the migration rate could be reduced from 59 to 11 % by using these stents, which still remains very effective.

The double-type metallic stent (Taewoong Medical, Seoul, Korea) is designed with an outer uncovered wire construction that allows the meshes embedding themselves in the esophageal wall, in the purpose of preventing migration (Fig. 1). It is associated with an inner silicone membrane in order to prevent a mucosal ingrowth. This stent has already been successfully used for malignant indications, but it has never been studied in benign indications, although it might probably both improve the endoscopic withdrawal and reduce the migration rate.

Fig. 1
figure 1

Double-type metallic stent: external view of the released stent with an outer uncovered wire construction to allow the mesh to embed itself in the oesophageal wall

Full size image

Thus, the aim of our study was to evaluate retrospectively the efficacy of double-type metallic stents in the management of leakages or fistulas, perforations, in the upper GI tract.

Materials and methods

This retrospective study was conducted between 2010 and 2013. All the patients referred to our institution for anastomotic leakages or fistulas, perforations, after upper GI tract surgery were managed with this new endoscopic strategy. The diagnosis was made based on clinical symptoms (fever, respiratory distress, hemodynamic shock, and increased output of external drainage) combined with the findings of the CT Scan and/or a radiological esophago-gastro-duodenal follow-through examination that confirmed the leak. The same experienced operator performed all the endoscopic procedures using a large working-channel gastroscope (diameter 3.8 mm, Hoya-Pentax, Tokyo, Japan). They were carried out on patients under general anesthesia with endotracheal intubation and under fluoroscopic control.

The endoscopic procedure was performed on an X-ray table using a contrast agent to determine the type and the location of the upper GI tract complication, including fistula or anastomotic leakage (Fig. 2). A simple fistula was defined as having a single outlet and a complex fistula corresponded to a cavity with multiple outlets. The fistula was considered chronic when the delay between the surgery and endoscopic treatment was more than 90 days. Then, once the fistula has been characterized, the next steps were the diversion of the fistula with the double-type metallic stent (Taewoong Medical, Seoul, Korea). The placement procedure was performed using a Savary-Gilliard® guide wire (Cook Medical, Winston-Salem, NC) as tutor and under X-rays control since this is a no TTS stent. A systematic endoscopic and fluoroscopic control was finally carried out to assess the good position of the stent and the absence of residual leak using a contrast agent (Figs. 3 and 4). In all the cases, the stent was removed after 4 weeks according a “sock” technique from its distal extremity in order to avoid a mucosal impaction. In case of chronic fistula, an added treatment for closing the fistula’s primary orifice could be performed using the OTSC clipping system (OVESCO Endoscopy, Tubingen, Germany) (Fig. 5). An intravenous antiemetic medication was systematically administered (Ondansetron 8 mg and Chlorpromazine 25 mg) for 48 h, and a radiological esophago-gastro-duodenal follow-through study was performed after two post-operative days with Gastrografine® (Bayer Santé, Paris, France) to ensure the absence of leaks before refeeding the patients.

Fig. 2
figure 2

The different step of the procedure in a patient with chronic oesotracheal fistulas: radiological assessment of the oesotracheal fistula

Full size image
Fig. 3
figure 3

The different step of the procedure in a patient with chronic oesotracheal fistulas: endoscopic control of the good position of the stent

Full size image
Fig. 4
figure 4

The different step of the procedure in a patient with chronic oesotracheal fistulas: fluoroscopic control showing both the OTSC and the DTMS device bridging the fistula

Full size image
Fig. 5
figure 5

The different step of the procedure in a patient with chronic oesotracheal fistulas: suture of the primary orifice with the Ovesco clip

Full size image

The primary endpoint was the efficacy of the double-type stent for the treatment of anastomotic leakages or fistulas in the upper GI tract. The secondary endpoint was to determine the ability for removing the DTMS at the end of the treatment. A successful treatment was defined by a perfect seal of the fistula at the end of the endoscopic management. That was confirmed by the complete disappearance of the septic symptoms associated with a satisfying radiological and endoscopic examination after the removal of the stent. A primary success was defined by the complete closure of the leakage after only one intervention (one stent placement). The secondary success corresponded to a recovery that required one or more additional treatments with the same double-type metallic stent. The qualifying period to assess the efficacy of the stent for sealing the leak or the fistula was three months.

Results

Characteristics of the patients: (Table 1)

Thirty-six patients underwent an endoscopic double-type metallic stent placement for benign indications between 2010 and 2013 in our institution: twenty-four patients had a post-operative fistula, among them fifteen after Sleeve Gastrectomies and the remaining after other kind of upper GI surgeries. Eight had an anastomotic leakage, and four had an esophageal perforation; two were Boerhaave syndromes, and two occured during endoscopy. The mean age of the patients was 56 [25–83] years old with a predominance of women (59 %). The median delay of endoscopic treatment was 23 [1–720] days. Ten patients (27 %) fulfilled the chronic fistula criteria.

Table 1 Characteristics of the patients
Full size table

Anterior management

Seventeen patients (47 %) were previously unsuccessfully treated with at least one fully covered metallic stent. A double-type metallic stent could be successfully placed in all the patients and was removed after a median time of 32 [20–71] days. A median number of 3 [2–11] endoscopies had been performed since the placement of the first DTMS. Fourteen patients (39 %) received a combined treatment with OTSC placement, including three patients with a chronic and fibrotic fistula that needed two OTSCs.

Efficacy: (Table 2)

A final complete healing was achieved in 26/36 patients included (72 %): in sixteen among the twenty-four cases (66.6 %) of fistulas, including twelve on the fifteen (80 %) patients with a postsleeve gastrectomy fistula and five on the nine patients (55.5 %) with other etiologies of fistula, in six on the eight (80 %) patients with anastomotic leakages, and in three on the four with an esophageal perforation. A primary success was achieved in 58 % of the cases (twenty-one patients). The success rate in the sub-group of patients who underwent a combined treatment with OTSC was 78.5 % (11/14). In the sub-group of patients with a chronic fistula, the success rate was 77.8 % (7/9) whereas it was 70.4 % (19/27) in the sub-group of acute fistulas. If a previous endoscopic treatment had been applied, the success rate was 65 % (11/17) against 79 % without any prior treatment (15/19).

Table 2 Primary, secondary, and overall efficiency of DTMS, with or without associated treatment
Full size table

Five patients (14 %) underwent a second stenting session, three of them with a DTMS, one with a full covered metallic stent, and the last one with both kinds of stents. All the patients with additional stenting treatment achieved a secondary success after 2–3 months. Finally, ten patients (28 %) had a complete failure: five underwent a surgery, three, who had a chronic fistula, were treated with a double pigtails stents, and two with the OTSC device alone.

Complications

All the stents were removed successfully without any adverse events. The spontaneous migration rate of the DTMS was 16.6 % (6 cases). There were only three minor adverse events that consisted in esophageal strictures at the end of the endoscopic treatment. All of them were successfully treated by endoscopic balloon dilation without recurrence. There was one death during the study period not related to the stent but consecutive to a tracheoesophageal fistula. This patient had been in intensive care unit since the complication occured, which had lead to a spetic shock associated with mediastinitis, pulmonary over infections, and refractory acute respiratory distress. He finally died because of this very grave clinical condition, in spite of the stent placement and the closure of the fistula.

Discussion

Most of the stents usually used for the management of fistulas or leakages in benign indications are fully covered in order to optimize the removability. It results in a high migration rate with a loss of efficiency and the recurrence of sepsis, which is still a life-threatening event in intensive care unit. Deviere’s team used partially covered stent to decrease the migration rate up to 11 % [9, 10], but the ablation of such stent requires the “stent in stent” technique to induce necrosis of the ingrowths tissues at the tip of the stents, thus facilitating the stent removability. Since the DTMS combined both inner-covered part and outer uncovered part, we decided to use this stent in benign indications, although it was initially dedicated for malignancy. We assessed the efficacy as well as the migration rate and the removability, and we compared our results to those of previous major series.

Thus, in a recent systematic review and meta-analysis, the use of self-expandable stents (SESs) was a minimally invasive, safe, and effective alternative in the treatment of leaks complicating bariatric surgery [11]. A total of 189 relevant articles were reviewed of which 7 studies (67 patients with leaks) met with the inclusion criteria. The pooled proportion of successful leak closures by using SESs was 87.77 % [95 % CI 79.39–94.19], the pooled proportion of successful endoscopic stent removals was 91.57 % [95 % CI 84.22–96.77], and the stent migration rate was noted in 16.94 % [95 % CI 9.32–26.27]. They concluded that the use of self-expandable stents could minimize the need for surgical revision and improve the patient outcomes.

In the experience of Yimcharoen et al. [12], eighteen patients (12 women, mean age 51 ± 15 y.o.) underwent an endoscopic stent placement for anastomotic complications among which fourteen were bariatric patients. A total of 31 stents of different kinds were used to treat anastomotic leaks (n = 13), strictures (n = 3), and fistulas (n = 2). This treatment was completely successful in managing anastomotic complications in 72 % of the cases (n = 13). A stent migration occurred in four cases and was amenable to endoscopic management. Two patients died, both unrelated to stent placement.

Swinnen et al. [10] reported their experience in using self-expanding metal and plastic stent in patients with benign upper GI leaks or perforations. Eighty-eight patients were treated and followed up in this study. All the self-expanding metal stents (n = 153) were successfully placed and could be removed in 96 % of the cases. The healing of the leaks and perforations was achieved in 77.6 % of cases with one single endoscopic session, and in 84.2 % of cases (64 patients) after a prolonged or repeated endoscopic treatment. Spontaneous migration occurred in only 11.1 % of the cases. They encountered minor adverse events (dysphagia, hyperplasia, and rupture of coating) in 20.9 % and major adverse events (bleeding, perforation, and tracheal compression) in 5.9 %.

Finally, Van Boeckel and al. [13] treated a total of fifty-two patients with leakages or perforations using 83 esophageal stents (61 partially covered self-expanding metal stent, 15 fully covered self-expanding metal stent, and 7 self-expanding plastic stent). The endoscopic stent removal was successful in 86 % of the cases (except for the PCSEMS). The clinical overall success was achieved in 34 patients (76 %, intention-to-treat: 65 %) after a median number of one stent and a median stenting time of 39 days. The migration rate was 19 %, and the complication rate was 46 % (tissue ingrowth, ruptured stent cover, food obstruction, pain, esophageal rupture, hemorrhage) and one (2 %) patient died of a stent-related cause.

In our previous experience, a combined endoscopic management associating OTSC clips and stenting should be considered for all patients presenting fistulas after sleeve gastrectomy [14]. In this series, this combined treatment was feasible and safe, and improved the global efficacy of endoscopic treatment with a complete removability. There was no complication at the removal and no migration of the double-type metallic stent when it was used.

In this current study, we evaluated the new double-type metallic stent which could combine the advantages of non-covered and fully covered stents on terms of migration and removability. In spite of the small sample size, our results were promising, since the migration rate in our series was 16.6 % that is comparable with the other pooled studies using partially covered stents. The removability was excellent as well and possible in all the cases using a “socks” technique. As concerned the primary endpoint, the effectiveness (72 %) was consistent with the other studies published (Table 3) in the last past years with various kinds of stents. Some patients needed several stent placements to achieve a success in sealing fistulas or leakages. Currently, as efficacy between stents is not different, stent choice should depend on expected risks of stent migration and tissue in- or overgrowth. In our study, they were always removable despite the external uncovered part, which allowed a low migration rate as well. There was no stent tissue obstruction due to an ingrowth during the follow-up, and there were only three stenotic complications at the end of the endoscopic management, all successfully dilated. No deaths or severe adverse events related to the stent placement were observed during the study period. However, even if the double-type metallic stent sounds a really promising approach in this new indication, we acknowledge that further studies are required to confirm these outcomes. Nevertheless, those stents represent a new and very interesting alternative for the management of anastomotic complications, especially in the upper GI tract and whatever the type of surgery.

Table 3 Studies comparison
Full size table

In conclusion, this type of stent is a new and efficient way to treat post-operative fistula or leakage of the upper GI tract, allowing a low migration rate and an excellent removability.