Closure of benign leaks, perforations, and fistulas with temporary placement of fully covered metal stents: a retrospective analysis.

ORIGINAL ARTICLE

Closure of Benign Leaks, Perforations, and Fistulas With Temporary Placement of Fully Covered Metal Stents: A Retrospective Analysis Aitor Orive-Calzada, MD,*w Ańgel Calderoń-Garcıá, MD,z Antonio Bernal-Martıńez, MD,* Ana B. Dıáz-Roca, MD,z Irantzu Barrio-Beraza, MSc,y Jose L. Cabriada-Nunõ, MD,* and Victor M. Orive-Cura, MDwz

Introduction: Partially covered self-expanding metal stents (SEMS), have been suggested as an alternative to surgery in the treatment of esophageal fistulas of benign etiology. Nevertheless, uncomplicated removal remains difficult. The use of fully covered (FC) SEMSs could solve this problem. Objectives: To review our experience with FC-SEMS placement in patients with benign upper gastrointestinal leaks or perforations. We wanted to assess successful closure of the perforations and short-term and long-term complications. Materials and Methods: Multicenter study, including 3 tertiary centers. Retrospective review of patients who underwent FC-SEMS placement for benign perforations. Results: Eighty-eight stents were placed in 56 patients. We achieved leak closure in 44 patients (78.6%). There were 18 migrations. All of them could be solved endoscopically. A severe septic situation was associated with a higher mortality rate (27.6% vs. 7.4%; P = 0.049) and a lower success rate (34.5% vs. 7.4%; P = 0.088), compared with those patients who did not present severe sepsis. However, these differences could not be confirmed by multivariable analysis. The results in the subgroup of 11 patients with leaks after sleeve gastrectomy were also good (73% success without surgery and 0% mortality). Conclusions: Temporary placement of FC-SEMS for benign perforations, fistulas, and leaks is feasible in sealing the leaks. All migrations could be solved endoscopically. It is very important to insert the stent before sepsis is established. This article also would be an addition to the growing body of literature supporting stenting as a good alternative if not standard approach to controlling these leaks. Key Words: esophageal stenting, leak, perforation, fistula

(Surg Laparosc Endosc Percutan Tech 2014;24:528–536)

P

erforation of the esophagus is a challenging and lifethreatening complication associated with significant morbidity and mortality rates. Its 2 main causes are

Received for publication January 19, 2013; accepted March 24, 2013. From the *Department of Gastroenterology and Hepatology, Endoscopy Unit, Galdakao-Usansolo Hospital, Galdakao, Bizkaia; wEndoscopy Unit, Clı´ nica Dr V. San Sebastiań; zDepartment of Gastroenterology and Hepatology, Endoscopy Unit, Basurto Hospital, Bilbao, Bizkaia; and yDepartment of Applied Mathematics, Statistics and Operational Research, Universidad del Pais Vasco, UPV/EHU, Leiona, Spain. Institutions participating in the study: Galdakao-Usansolo Hospital, Basurto Hospital, Clı´ nica Dr V. San Sebastiań. The authors declare no conflicts of interest. Reprints: Aitor Orive-Calzada, MD, Department of Gastroenterology, Endoscopy Unit, Galdakao-Usansolo Hospital, B1 Labeaga s/n 48960, Galdakao, Bizkaia, Spain (e-mail: [email protected]). Copyright r 2014 by Lippincott Williams & Wilkins

528 | www.surgical-laparoscopy.com

anastomotic leakage and endoscopy-related perforation. The management of a patient with an upper gastrointestinal (GI) leak is complex and often involves treatment of sepsis, organ failure, and nutritional deficits, in addition to treating the underlying leak. So far, the standard management of esophageal perforation has traditionally been the surgical treatment. However, it is accepted that patients with limited contamination, without manifestations of evolving sepsis and organ failure, can be treated safely with simple drainage and antibiotics.1 The management of larger defects, especially when associated with significant contamination, tissue loss, and systemic sepsis is more difficult. In the setting of large anastomotic leaks, surgical reexploration is recommended, because conservative treatment is mostly unsuccessful.2–4 In spite of current advances, mortality from clinically apparent thoracic leakage remains high with these traditional management options, commonly approaching 50%.5–11 Self-expanding metal stents (SEMSs) are indicated for the palliation of symptoms associated with malignant esophageal obstruction.12,13 Various types of covered SEMS are currently used in the treatment of malignant esophageal fistulas with excellent results and good short-term quality of life, this treatment being considered the standard approach to malignant perforation and fistulas.8,14–16 The use of these SEMS is a theoretically attractive approach to benign esophageal perforations, but previous reports about SEMSs placed for benign esophageal diseases speculate on the possibility that the risk of complications and potential lack of removability could be high.17,18 These difficulties to remove the stents usually happen with partially covered (PC)-SEMS because hyperplasia on the uncovered parts at both ends can make those SEMS difficult to remove.19–22 Some authors advocate the use of a self-expanding plastic stent (SEPS) inside the SEMS to facilitate removal by inducing pressure necrosis of the hyperplasia, especially after >6 weeks of treatment.23 Fully covered (FC)-SEMS have been developed to address these problems, being easily removable even after 6 weeks of treatment. In contrast, and because hyperplastic tissue is not usually present in those FC-SEMS, they could theoretically present a higher migration rate.23 In this study, we describe the results of 56 consecutive patients who underwent FC-SEMS placement for upper GI perforation at 3 different institutions between August 2004 and November 2010. Factors influencing the success and complications of this approach are discussed. Our primary endpoint was to assess the efficacy and complication rates of FC-SEMS for treatment of upper GI leak or perforation. As secondary endpoints we wanted to asses if there was any

Surg Laparosc Endosc Percutan Tech

Volume 24, Number 6, December 2014



isolated variable that could contribute to the success or to the failure of the treatment.

PATIENTS AND METHODS Patients The clinical records of all adult patients who were treated for benign superior digestive tract (mostly esophageal) fistulas and perforations with FC-SEMS at 3 different institutions [Basurto Hospital (HBAS), Galdakao-Usansolo Hospital (HGU), and Dr V. San Sebastian Clinic (CVSS)] between August 2004 and November 2010 were retrospectively reviewed. Fifty-six patients with benign superior digestive tract fistulas and perforations were treated with stents in our 3 centers during the study period. There were a total of 88 SEMS placements in these patients. We tried to detect if there was any isolated variable that could contribute to the success or to the failure of the treatment. The patients were divided in 2 groups: 1 with clinically successful results and the other one with nonsuccessful results. Successful results were defined as endoscopical, radiologic (by contrast esophagography) and clinical evidence of perforation closure. We also analyzed if there was any factor related to a higher mortality derived from this treatment. We graded our patients in 4 groups [(a): patients without sepsis; (b): patients with nonsevere sepsis; (c): patients with severe sepsis; (d): patients with multiorgan failure] following the definitions of sepsis established at the International Sepsis Definitions Conference in 200124 (Table 1). Indications were (a): postsurgical fistula in 44 patients (78%), subdivided into fistula after bariatric surgery (being all of them sleeve gastrectomies) in 11 patients (25%), total gastrectomy in 15 patients (34%), esofagectomy in 11 patients (25%), Nissen in 5 patients (11.4%), and others in 2 patients (4.5%); (b): perforations caused by endoscopic treatments in 6 cases (11%); (c): Boerhaave syndrome in 4 cases (7.14%); (d): in the remaining 2 cases (4%) perforations were caused by a wound to the esophagus caused by a fish bone in 1 case and severe ulcerative esophageal reflux in the other. Patients were identified through a previously prospectively filled database that was used in the 3 centers. This prospectively recorded database included data about patients’ demographics and parameters such as etiology; delay until FC-SEMS placement (after the procedure that

Fully Covered Stents for Closure of Benign and Fistulas

caused the leak or the perforation); clinical situation at the time of SEMS placement: presence or absence of severe sepsis criteria; stent size and location; duration of stent permanence; evidence of stent migration; number of days a FC-SEMS was left in place; technical success with stent placement and removal; short-term (1 mo after stent removal) clinical outcomes, mortality, and stent-related complications. These results were then reviewed by the same 2 physicians (V.M.O.C. and A.O.C.) and additional and missed data were subsequently retrospectively retrieved from the institutional medical record of each patient and included data about long-term results complications. Patients with known upper GI malignancy were excluded from the study, except patients who underwent surgery for a curable upper GI malignancy with clean resection margins. None of the patients had previous history of radiation.

Stenting Protocol A multidisciplinary team of surgeons and gastroenterologists evaluated patients for endoluminal stenting. During the consent process, patients were informed that these stents are not currently designed neither used for closure of esophageal leakages. Informed consent was obtained for use of the stent in all the patients. Small perforations were measured endoscopically, by close apposition of an endoscopic forceps of a determined size. Bigger perforations were measured by calculating their length according to distance from the teeth to the distal and proximal ends of the perforation. All the FC-SEMS were placed endoscopically, with or without using fluoroscopy, with a guidewire-assisted deployment. We only used fluoroscopical aid when it was technically difficult to place the stent with just an endoscopical technique (ie, total disruption of the anastomosis). Occlusion of the leak was confirmed within 24 hours of placement with a water-soluble contrast study. Methylene blue ingestion was also used in doubtful cases (Figs. 1–4). Every patient was sedated with propofol by anesthesiologists or by doctors from intensive care unit staff. All the relevant mediastinal abscesses and collections were drained surgically or radiologically if it was possible. We did not consider endoscopic drainage of the collections in our patients. After the stent was placed, we used antiemetic drugs (ondansetron) if the patients presented nausea. Conventinal

TABLE 1. Sepsis’ Definitions Established by the International Sepsis Definitions Conference in 200124

Definitions Systemic inflammatory syndrome (SIRS) Body temperature Heart rate Respiratory rate White blood cell count Sepsis Severe sepsis Septic shock

r

2014 Lippincott Williams & Wilkins

Defined by the presence of Z2 of the following findings < 361C (96.81F) or >381C (100.41F) > 90 beats/min > 20 breaths/min or, on blood gas a PaCO2 < 32 mm Hg < 4000 cells/mm3 or >12000 cells/mm3 or >10% band forms (immature white blood cells) Defined as SIRS in response to a confirmed infectious process Defined as sepsis with organ dysfunction, hypoperfusion, or hypotension Defined as sepsis with refractory arterial hypotension or hypoperfusion abnormalities in spite of adequate fluid resuscitation

www.surgical-laparoscopy.com |

529

Orive-Calzada et al



removal, placement of a new stent for 6 more weeks was considered. If fistula was not found during endoscopy at the stent retrieval, we performed a hydrosoluble-contrast esophagography to document fistula closure. Resolution or worsening of the mediastinal and/or pleural contamination was monitored by clinical, analytic, and radiologic (by computed tomography scan) controls. All patients were treated with a high dose of a proton pump inhibitor and with broad-spectrum antibiotics.

Definitions

FIGURE 1. Esophageal perforation after dilation of achalasia. Endoscopic image.

analgesics (paracetamol and metamizole) were given if the patient had pain. We used FC-SEMS in all patients. In 1 patient, after presenting several migrations, we used a PC-SEMS (Evolution Controlled Release Stent; Cook Medical, Winston Salem, NC) using a SEPS for trying to remove the latest. Almost all the FC-SEMS used were Hanarostent (Hanarostent covered esophageal stent, MI Tech; IZASA, Seoul) but one SX-ELLA (SX-ELLA esophageal HV stent; Ella-CS, Hradec Kra´love´, Czech Republic) was used in a single patient. The length and central diameter of stent chosen varied depending on the presence or absence of anatomic barriers to stent migration (ie, anastomotic strictures, presence or absence of a competent cardias, y) and depending on the size and location of the leak. These FC stents are designed to be removable by pulling a drawstring (they present 1 at the proximal end and another 1 at the distal end) with forceps. When one of these drawstrings is pulled by forceps, the stent collapses, allowing it to be retrieved atraumatically. We used this technique to remove all the FC-SEMS. The stents were removed endoscopically 6 weeks after their placement. If fistula was still present after stent

Stent migration was defined as a radiographic or endoscopic visualization of the stent in a position different from where it was originally placed. Clinical outcomes were evaluated by review of available clinical, endoscopic, and radiographic studies for each patient around the time of stent removal. Clinical success was defined as the resolution of the fistula documented in both radiologic (by contrast esophagography), and endoscopic studies in an asymptomatic patient. We defined the delay of SEMS placement after the index procedure as immediate (< 24 h), rapid (1 d to 1 wk), moderate (1 wk to 1 mo), and late (> 1 mo), as previously described by Swinnen et al.23

Statistics All analyses were performed by using statistical software (SPSS 11.0; SPSS Inc., Chicago, IL). Descriptive analyses were performed to compare patients who succeeded or patients who failure to the treatment, so as to compare patients who died during the study period. Frequency and percentages for categorical variables and mean and SDs for continuous variables were computed. Differences between groups were evaluated using the w2 test or Fisher exact test when >25% of expected frequencies fell 3 cm) versus small openings (r3 cm). Variables for which P-value 3 cm) fistulous openings appeared to be more likely to be unsuccessful, but differences were not statistically significant (r3 cm: 8/48: 16.7% vs. >3 cm: 4/8: 50%; P = 0.055). However, these differences were not confirmed in the multivariate analysis because of the correlation between these variables and the ASA type variable. Furthermore, patients with bigger fistulous openings appeared to have higher ASA scores than those with small fistulous opening but these differences were not statistically confirmed (35.7% vs. 17.1%; P = 0.1448).

Surviving and Nonsurviving Groups Nine patients (16.1%) died with a stent in place or within a month’s period after stent extraction. Three of these deaths (5.3%) were produced by etiologies or circumstances not related to the fistula neither to stents: 1 by a cerebrovascular accident, 1 by a nosocomial pneumonia, and 1 by a neoplasia. The other 6 deaths (10.7%) were secondary to septic processes. All these 6 patients who died from sepsis or mediastinitis, already suffered from severe sepsis before stent insertion and, in at least 3 of them, collections could not be treated aggressively with drainages before stent insertion because of their bad clinical situation. In one of these patients who died from mediastinitis, surgery had already failed before trying stenting (Tables 3 and 4). When analyzed, baseline characteristics between the surviving and nonsurviving groups, in the univariate analysis, we found that there were no statistically significant differences in sex [(28 (59.6%)#/19 (40.4%)~ vs. 8 (88.8%)#/1 (11.1%)~ (P = 0.1359)], site of the fistulous opening [proximal third: 6 (12.76%) vs. 1 (11.1%); mid third: 3 (6.4%) vs. 0 (0%); distal third 38 (80.8%) vs. 8 (88.8%); P = 0.788], failure of a previous surgical attempt to repair the perforation [18 (38.3%) vs. 1 (11.1%); r





TABLE 3. Differences in Baseline Characteristics Between the Success and Nonsuccess Groups and Between the Surviving and Nonsurviving Groups

Success Group

Characteristics

Nonsuccess Group

N 44 (78.57%) 12 (21.42%) Age (mean ± SD) 60.74 ± 13.886 61.09 ± 2.118 Sex (M:F) Male 26 (59.1%) 10 (83.3%) Female 18 (40.9%) 2 (16.6%) Diameter of the fistulous opening (mean ± SD) 1.79 ± 1.273 2.68 ± 1.167 Previously diagnosed severe sepsis due to the perforation Yes 19 (43.2%) 10 (83.3%) No 25 (56.8%) 2 (16.7%) Delay of SEMS placement after the index procedure 10.88 ± 8.247 22.18 ± 33.4 ASA score (I-III/IV) I-III 38 (86.4%) 6 (50%) IV 6 (13.6%) 6 (50%) Etiology Medical etiology 8 (18.2%) 2 (16.6%) Postsurgical 36 (81.8%) 10 (83.3%) B-II 1 (2.8%) 0 Bariatric surgery 8 (22.2%) 3 (30%) Esophagectomy 8 (22.2%) 3 (30%) Total gastectomy 12 (33.3%) 3 (30%) Nissen 5 (13.9%) 1 (10%) Piloroplasty 1 (2.8%) 0 Esophagomiotomy 1 (2.8%) 0 Big fistulous opening (> 3 cm) r3 cm 40 (90.9%) 8 (66.7%) > 3 cm 4 (9.1%) 4 (33.3%)

P 0.991 0.178

Surviving Group

Nonsurviving Group

47 (83.9%) 59.17 ± 14.49

9 (16.1%) 71.33 ± 11.32

28 (59.6%) 19 (40.4%) 0.018 1.83 ± 1.23 0.014

8 (88.8) 1 (11.1%) 2.83 ± 1.12

P 0.0162 0.1359 0.0176 0.088

22 (46.8%) 7 (77.8%) 25 (53.2%) 2 (22.2%) 0.236 13.71 ± 18.283 10.56 ± 6.589 0.013 39 (83%) 5 (55.5%) 8 (17%) 4 (44.4%). 0.425 7 (14.3%) 2 (22.2%) 39 (79.6%) 7 (77.8%) 1 (2.6%) 0 11 (28.2%) 0 8 (20.5%) 3 (42.9%) 11 (28.2%) 4 (57.1%) 6 (15.4%) 0 1 (2.6%) 0 1 (100%) 0 0.055 42 (89.4%) 5 (55.5%) 4 (8.5%) 4 (44.4%)

0.944 0.0869 0.529

0.0185

Values in bold indicate statistically significant differences. SEMS indicates self-expanding metal stents.

P = 0.1145], delay of SEMS placement after the index procedure (13.7 ± 18.2 vs. 10.6 ± 6.6; P = 0.944), and etiology (Table 3). We found that patients from nonsurviving group had a trend to present higher ASA scores, but these differences were neither statistically significant [ASA I to III 5/44 (11.4%) vs. ASA IV 4/13 (30.8%); P = 0.087]. Therefore, an initial severe septic situation was associated with a statistically significant higher mortality rate (8/29: 27.6% vs. 2/25: 7.4%; P = 0.049) and with an also statistically significantly lower success rate in achieving the closure of the fistulous opening (10/29: 34.5% vs. 2/27: 7.4%; P = 0.021), compared with those patients who did

not present this severe sepsis situation, the differences being statistically significant. However, we found a statistically significant difference in age, the surviving patients being significantly younger than those who died (59.1 ± 14.6 vs. 79.3 ± 11.17; P = 0.027) (Table 3). There was also a significant difference with regard to the size of the fistulous opening, presenting patients from the nonsurviving group higher fistulous openings (2.83 ± 1.118 vs. 1.8 ± 1.267; P = 0.015). Either the age and the size of the fistulous openings were confirmed to be statistically significant in a multivariate analysis (Table 5) with and area under the receiver operating characteristic curve value of 0.86.

TABLE 4. Complication and Mortality Rates

None No. Mild No. Severe No. (% Per Patient) (% Per Patient) (% Per Patient) Complications Migration Perforation Atrial fibrilation during procedure Death with stent placed or 1 mo after extraction CVA Nosocomial pneumonia Neoplasia Sepsis

34 38 53 55

(61%) (67.8%) (94.6%) (98.2%)

19 (33.9%) 18 (32.1%) 0 1 (1.8%)

3 (5.3%) 0 3 (5.3%) 0

Total Per Patient No. (%) n = 56 22 18 3 1 9 1 1 1 6

(39%) (32.1%) (5.3%) (1.8%) (16%) (1.8%) (1.8%) (1.5%) (10.7%)

Per Stent No. (%) n = 88 22 18 3 1 9 1 1 1 6

(4%) (20.45%) (3.4%) (1.1%) (10.2%) (1.1%) (1.1%) (1.1%) (6.8%)

CVA indicates cerebrovascular accident.

r



533


Orive-Calzada et al


TABLE 5. Multivariate Analysis: Multivariate Logistic Regression Model to Explain Death Among Patients With Esophageal Fistula

Parameters Age Size of the fistulous opening >3 cms (yes vs. no)

Estimate

P

OR

95% CI

0.1 2.89

0.0278 0.0125

1.105 17.94

1.01-1.21 1.86-172.94

The area under the curve (AUC) = 0.86. CI indicates confidence interval; OR, odds ratio.

This way, the more severely ill patients, with more comorbidities, with ASA IV score, were less likely to seal their fistulous openings and were also less likely to survive. We were successful in sealing the perforations in 13/16 (81.25%) of patients in whom a previous surgery had already failed to solve the fistulous opening. A total of 22 patients (39%) suffered from some kind of complication, most of them being mild (19/22: 86.36%). From these mild complications, most were migrations (18/ 19: 94.74%), all of them solved endoscopically and the remaining mild complication was 1 patient (1/19: 5.26%) with atrial fibrillation related to sedation which was pharmacologically reverted. There were only 3/56 (5.35%) severe complications, all of them related to perforations due to the stent. Two of the 3 perforations were solved with surgery, whereas the third one was solved by inserting a second, larger stent. Most of the patients had sepsis criteria at the time of stent insertion, being severe sepsis criteria in 29 of 56 patients (51.78%) of the patients before stent insertion. All 6 patients (100%) who died from mediastinitis and severe sepsis, fulfilled severe sepsis criteria before stent insertion, whereas there were no deaths from mediastinitis in the group who did not present severe sepsis criteria before stent insertion; these differences being statistically significant (P = 0.024).

DISCUSSION To our knowledge, this is the world’s largest series of cases where FC-SEMS were used for the treatment of benign upper GI leaks and perforations. Our report describes 56 consecutive patients with nonmalignant, traumatic perforations of the esophagus that were sealed by using covered metallic stents. As previously mentioned, traumatic esophageal perforation is a serious injury with high morbidity and mortality if untreated.25 Iatrogenic perforation of the esophagus after instrumentation or surgery is the most common cause, but perforation may also occur spontaneously during vomiting (Boerhaave syndrome).26 Successful management depends on early diagnosis and prompt treatment. However, after surgery, a diagnosis of perforation is often difficult to establish, and, in other cases, diagnosis depends on the physician’s alertness.27 The classic treatment options include surgical repair, esophagectomy, or cervical exclusion. Primary surgical closure and mediastinal drainage within 24 hours of the injury have shown to improve survival rates.28 However, if the diagnosis is not early enough, surgery carries high morbidity and mortality, particularly in patients with mediastinal and pleural contamination. These patients may be better treated conservatively, because mortality from a surgical intervention equals that of a conservative approach.29


Covered metallic stents have been successfully used to alleviate patients with malignant perforation30 and fistula,31 and are also reported to have been useful in individual cases of Boerhaave syndrome.32–38 In our series, primary closure of the fistulous opening was achieved in 79% of patients. This result is similar to that of other series of PC-SEMS or FC-SEMS placements.9,10,23,39,40 The death rate in our series is 16%, where 3 of these deaths are not related to the technique neither to the fistula. So the real mortality rate is 10.7%. This mortality rate is much lower than that reported from the classic series of surgical treatment.6–8,41–43 There is another large sample size paper published by van Boeckel et al.44 In van Boeckel and colleague’s paper, they analyze their experience with temporary stenting for nonmalignant esophageal ruptures and anastomotic leaks in 52 patients also with good results (success 76%, complications 46%, mortality rate 10%). However, van Boeckel and colleagues used a great variety of stents including FC-SEMS (15 stents), PC-SEMS (61 stents), and SEPS (7 stents). So, they used only 15 FC-SEMS in this paper. The heterogeneity in the kind of stent used is the main complaint in this paper. Moreover, endoscopic stent removal was not successful in 8 of the patients of the PCSEMS group because of tissue ingrowth and/or overgrowth, requiring esofagectomy in 1 case and a rupture occurring during stent removal in other 2 cases.44 Recently, David and colleagues have published a paper with similar results (clinical success 76.6%, complications 40%, mortality 10%), evaluating the results of SEMS (FC-SEMS and PC-SEMS) in closing the leakages in 30 patients with esophageal or gastric perforation and intrathoracic contamination, who were treated with cSEMS.45 They only analyzed the results of this subgroup of patients where intrathoracic contamination was found. There is another large Belgian series recently published by Swinnen et al,23 where they use PC-SEMS for the treatment of benign leaks and perforations in 88 patients with excellent results (success 72.7%, complications 51.1%, death rate 7%). In Swinnen and colleague’s paper, the authors often needed complex and repeated endoscopic maneuvers to remove the prosthesis. In case of severe hyperplastic development, they required frequent extra sessions to insert a plastic stent (SEPS) inside the PC-SEMS for 10 days, in order to induce a pressure necrosis of the hyperplastic tissue that could facilitate the removal of the SEMS. They used these PC stents in order to avoid migration. However, in our series, we used FC-SEMS, instead of the PC-SEMS that were used by Swinnen and colleagues. We could easily remove all the FC-SEMS and all the migrations could be solved endoscopically, those being considered a minor complication. We think that our procedure is much simpler and probably cheaper, so it is not worth complicating so much the procedure to avoid r




migration. In addition, the only SEMS that we were unable to remove was the case of 1 patient who presented previous migration of several FC-SEMS, so we proceeded to insert a PC-SEMS. We could not remove that stent because of the development of hyperplastic tissue, even after inserting a SEPS inside the SEMS as described before by Swinnen et al.23 We think that is important to highlight the results obtained in the subgroup of patients with fistulas of bariatric surgery. We have included 11 patients in this subgroup. All these patients had fistulas after a sleeve gastrectomy. The leaks are one of the most important complications after a sleeve gastrectomy affecting to 2.2% of these patients.46 This complication usually present a high mortality (9%)47 being its treatment challenging. In our series, we have treated 11 fistulas after sleeve gastrectomy with good results (73% success without surgery and 0% mortality). This way, we think that our article would be an addition to the growing body of literature supporting stenting as a good alternative if not standard approach to controlling these leaks. The procedure we describe is an effective treatment even in those patients whose previous surgical treatments for closing the fistula had failed. We were able to achieve the closure of the fistulous opening in 13/16 (81.25%) patients where a second surgery had already failed for this setting. Although most of our explorations took place without the use of radiologic control, we succeeded in the insertion of all stents. This simplicity makes it a very attractive treatment for patients with severe or postoperative leaks, as can be performed at the bedside in intensive care unit. Although it is important to insert the stent before the establishment of the sepsis, precocity in the placement of the prosthesis is not essential if the patient is not septic. In our series, a severe septic situation is associated with a higher mortality rate (7/27: 24.14%) and with a lower success rate (19/29: 65.5%) in achieving the closure of the fistulous opening, compared with those patients who did not present a severe septic situation (1/27: 7.4% mortality rate: P = 0.088 and 25/ 27: 92.6% success rate: P = 0.016). However, these results were not confirmed by the multivariate analysis. Although patients with larger fistulous openings usually have a lower overall success rate, in a non-negligible percentage of patients with large fistulas (including 2 with openings of more than two thirds of the circumference), we had successful results. In addition, it should be noted that, despite the success rate in patients with large openings (> 3 cm), the difference in success may be because these patients had a greater tendency to be septic at the time of placement of the prosthesis. In conclusion, the use of FC-SEMSs for the treatment of upper GI leaks and perforations is feasible, relatively safe, and effective when it is combined with adequate drainage of the thoracic cavity, even in cases of severely ill patients or even if surgical treatment for fistula closure has failed before. This is a retrospective study with a highly selected patient population, which can negatively impact its veracity because of possible bias of selection or information. However, it is necessary to perform further comparative prospective studies comparing surgical and endoscopic approaches to confirm our results. ACKNOWLEDGMENTS The authors thank the Research Committee of the Galdakao-Usansolo Hospital for the help in editing this article. r



REFERENCES 1. Csendes A, Burdiles P, Burgos AM, et al. Conservative management of anastomotic leaks after 557 open gastric bypasses. Obes Surg. 2005;15:1252–1256. 2. Urschel JD. Esophagogastrostomy anastomotic leaks complicating esophagectomy: a review. Am J Surg. 1995;169:634–640. 3. Schwartz RW, Strodel WE, Simpson WS, et al. Gastric bypass revision: lessons learned from 920 cases. Surgery. 1988;104: 806–812. 4. Griffin SM, Lamb PJ, Dresner SM, et al. Diagnosis and management of a mediastinal leak following radical oesophagectomy. Br J Surg. 2001;88:1346–1351. 5. Patil PK, Patel SG, Mistry RC, et al. Cancer of the esophagus: esophagogastric anastomotic leak—a retrospective study of predisposing factors. J Surg Oncol. 1992;49:163–167. 6. Viste A, Geir E, Soreide O. Stomach cancer: a prospective study of anastomotic failure following total gastrectomy. Acta Chir Scand. 1987;153:303–306. 7. Sauvanet A, Baltar J, Le Mee J, et al. Diagnosis and conservative management of intrathoracic leakage after oesophagectomy. Br J Surg. 1998;85:1446–1449. 8. Urschel JD. Esophagogastrostomy anastomotic leaks complicating esophagectomy: a review. Am J Surg. 1994;169:634–640. 9. Salminen P, Gullichsen R, Laine S. Use of self-expandable metal stents for the treatment of esophageal perforations and anastomotic leaks. Surg Endosc. 2009;23:1526–1530. 10. Fisher A, Thomusch O, Benz S, et al. Nonoperative treatment of 15 benign esophageal perforations with self-expandable covered metal stents. Ann Thorac Surg. 2006;81:467–472. 11. Rodella L, Laterza E, De Manzoni G, et al. Endoscopic clipping of anastomotic leakages in esophagogastric surgery. Endoscopy. 1998;30:453–456. 12. Baron TH. Expandable metal stents for the treatment of cancerous obstruction of the gastrointestinal tract. N Engl J Med. 2001;344:1681–1687. 13. Siersema PD. Treatment options for esophageal strictures. Nat Clin Pract Gastroenterol Hepatol. 2008;5:142–152. 14. Saxon RR, Morrison KE, Lakin PC, et al. Malignant esophageal obstruction and espophagorespiratory fistula: palliation with a polyethylene-covered Z stent. Radiology. 1997;202:349–354. 15. Morgan RA, Ellul JPM, Denton ERE, et al. Malignant esophageal fistulas and perforation: management with plastic covered metallic endoprosthesis. Radiology. 1997;204:527–532. 16. Nicholson AA, Royston CMS, Wedgwood K, et al. Palliation of malignant esophageal perforation and proximal esophageal malignant dysphagia with covered metal stents. Clin Radiol. 1995;50:11–14. 17. Sandha GS, Marcon NE. Expandable metal stents for benign esophageal obstruction. Gastrointest Endosc Clin North Am. 1999;9:437–446. 18. Wadhwa RP, Kozarek RA. Use of self-expandable metallic stents in benign GI diseases. Gastrointest Endosc. 2003;58: 207–212. 19. Watson A. Self-expanding metal esophageal endoprostheses: which is best? Eur J Gastroenterol Hepatol. 1998;10:363–365. 20. Shimi SM. Self-expanding metallic stents in the management of advanced esophageal cancer: a review. Semin Laparosc Surg. 2000;7:9–21. 21. Dormann AJ, Eisendrath P, Wigginghaus B, et al. Palliation of esophageal carcinoma with a new self-expanding plastic stent. Endoscopy. 2003;35:207–211. 22. Sabharwal T, Morales JP, Salter R, et al. Esophageal cancer: self-expanding metallic stents. Abdom Imaging. 2005;30: 456–464. 23. Swinnen J, Eisendrath P, Rigaux J, et al. Self-expandable metal stents for the treatment of benign upper GI leaks and perforations. Gastrointest Endosc. 2011;73:890–899. 24. Levy MM, Fink MP, Marshall JC, et al. 2001 CCM/ESICM/ ACCP/ATS/SIS International Sepsis Definitions Conference. Crit Care Med. 2003;31:1250–1256.


535

Orive-Calzada et al


25. Walker WS, Cameron EWS, Walbaum PR. Diagnosis and management of spontaneous transmural rupture of the oesophagus. Br J Surg. 1985;72:204–207. 26. Brauer RB, Liebermann-Meffert D, Stein HJ, et al. Boerhaave’s syndrome: analysis of the literature and report of 18 new cases. Dis Esophagus. 1997;10:64–68. 27. Sawyer R, Philips C, Vakil N. Short- and long-term outcome of esophageal perforation. Gastrointest Endosc. 1995;41: 130–134. 28. Zwischenberger JB, Savage C, Bidani A. Surgical aspects of esophageal disease. Perforation and caustic injury. Am J Respir Crit Care Med. 2001;164:1037–1040. 29. Tilanus HW, Bossuyt P, Schattenkerk ME, et al. Treatment of esophageal perforation: a multivariate analysis. Br J Surg. 1991;78:582–585. 30. Morgan RA, Ellul J, Denton ER, et al. Malignant esophageal fistulas and perforations: management with plastic-covered metallic endoprostheses. Radiology. 1997;204:527–532. 31. Siersema PD, Schrauwen SL, van Blankenstein M, et al. Selfexpanding metal stents for complicated and recurrent esophagogastric cancer. Gastrointest Endosc. 2001;54:579–586. 32. Adam A, Watkinson AF, Dussek J. Boerhaave’s syndrome: to treat or not to treat by means of insertion of a metallic stent. J Vasc Intervent Radiol. 1995;6:741–746. 33. Dumonceau JM, Deviere J, Cappello M, et al. Endoscopic treatment of Boerhaave’s syndrome. Gastrointest Endosc. 1996; 44:477–479. 34. Yuasa N, Hattori T, Kobayashi Y, et al. Treatment of spontaneous esophageal rupture with a covered self-expanding metal stent. Gastrointest Endosc. 1999;49:777–780. 35. Eubanks PJ, Hu E, Nguyen D, et al. Case of Boerhaave’s syndrome successfully treated with a self-expanding metallic stent. Gastrointest Endosc. 1999;49:780–783. 36. Davies AP, Vaughan R. Expanding mesh sent in the emergency treatment of Boerhaave’s syndrome. Ann Thorac Surg. 1999; 67:1482–1483. 37. Tsunoda S, Shimada Y, Watanabe G, et al. Covered metal stent treatment of a patient with spontaneous rupture of the esophagus. Dis Esophagus. 2001;14:254–257.



38. Chung MG, Kang DH, Park DK, et al. Successful treatment of Boerhaave’s syndrome with endoscopic insertion of a selfexpanding metallic stent: report of three cases and a review of the literature. Endoscopy. 2001;33:894–897. 39. Eisendrath P, Cremer M, Himpens J, et al. Endotherapy including temporary stenting of fistulas of the upper gastrointestinal tract after laparoscopic bariatric surgery. Endoscopy. 2007;39:625–630. 40. Salinas A, Baptista A, Santiago E, et al. Self-expandable metal stents to treat gastric leaks. Surg Obes Relat Dis. 2006;2: 570–572. 41. Junemann-Rodriguez M, Awan MY, Khan ZM, et al. Anastomotic leakage post-esofagogastrectomy for esophageal carcinoma: a retrospective analysis of predictive factors, management and influence on longterm survival in a high volume centre. Eur J Cardiothorac Surg. 2005;27:3–7. 42. Michelet P, D’Journo X-B, Roch A, et al. Perioperative risk factors for anastomotic leakage after esophaguectomy. Influence of thoracic epidural analgesia. Chest. 2005;128:3461–3466. 43. Ruiz de Adana JC, Ortega-Deballon P, Alonso-Garcia MT, et al. Morbidity and mortality due to esophagojejunal anastomotic fistula after total gastrectomy for gastric cancer. Cir Esp. 2001;70:3–5. 44. van Boeckel PGA, Dua KD, Weusten BLAM, et al. Fully covered self-expandable metal stents (SEMS), partially covered SEMS and self-expandable plastic stents for the treatment of benign esophageal ruptures and anastomotic leaks. BMC Gastroenterol. 2012;12:19. 45. David EA, Kim MP, Blackmon SH. Esophageal salvage with removable covered self-expanding metal stents in the setting of intrathoracic esophageal leakage. Am J Surg. 2011;202: 796–801. 46. Parikh M, Issa R, McCrillis A, et al. Surgical strategies that may decrease leak after laparoscopic sleeve gastrectomy: a systematic review and meta-analysis of 9991 cases. Ann Surg. 2013;257:231–237. 47. Sakran N, Goitein d, Raziel A, et al. Gastric leaks after sleeve gastectomy: a multicenter experience with 2834 patients. Surg Endosc. 2013;27:240–245.

r


Temporary placement of fully covered self-expandable metal stents for the treatment of benign biliary strictures.

Temporary placement of a fully covered self-expanding metal stent to allow therapeutic ERCP.

Fully versus partially covered self-expandable metal stents in benign esophageal strictures.

Iatrogenic pharyngoesophageal perforations treated with fully covered self-expandable metallic stents (with video).

Post-sphincterotomy bleeding: fully-covered metal stents for hemostasis.

The role of esophageal stents in the management of esophageal anastomotic leaks and benign esophageal perforations.

Endoscopic management of gastrointestinal perforations, leaks and fistulas.

Clinical image in gastroenterology. Benign esophageal perforation resolved through the placement of a fully covered self-expandable metal stent.

Safety and efficacy of a fully covered large-diameter self-expanding metal stent for the treatment of upper gastrointestinal perforations, anastomotic leaks, and fistula.

Difficult removal of fully covered self expandable metal stents (SEMS) for benign biliary strictures: the "SEMS in SEMS" technique.

Endoscopic clip closure of gastrointestinal perforations, fistulae, and leaks.

Fully Covered Metallic Stents for the Treatment of Benign Airway Stenosis.

Plastic stents or covered self-expandable metal stents for benign biliary strictures: same song, different verse?

Biliary stenting is not a prerequisite to endoscopic placement of duodenal covered self-expandable metal stents.

Treatment of esophagopleural fistulas using covered retrievable expandable metallic stents.

The effect of paclitaxel-eluting covered metal stents versus covered metal stents in a rabbit esophageal squamous carcinoma model.

Endosonography-guided drainage of malignant fluid collections using lumen-apposing, fully covered self-expanding metal stents.

Fully Covered Self-Expandable Metal Stents for Treatment of Malignant Biliary Strictures due to Pancreatic Carcinoma.

A US Multicenter Study of Safety and Efficacy of Fully Covered Self-Expandable Metallic Stents in Benign Extrahepatic Biliary Strictures.

Fully covered self-expanding metal stents for refractory anastomotic colorectal strictures.

Large diameter fully covered self-expanding metal stent placement for palliation of proximal malignant esophageal strictures.

Temporary placement of a fully covered self-expandable metal stent in the pancreatic duct for aiding extraction of large pancreatic duct stones: preliminary data.

Long-term outcomes of covered self-expandable metal stents for treating benign biliary strictures.

Endoscopic treatment of nonstricture-related benign biliary diseases using covered self-expandable metal stents.