Abstract
Introduction:
Consecutive to an increase in the number of bariatric surgeries worldwide, the number of articles reporting complications have also increased. The most dreadful bariatric complication is represented by the leak, and the endoscopic stent is still the standard treatment for some bariatric teams despite the poor quality of life and associated complications. The purpose of this review was to identify the very rare cases of aortic injuries associated with stent use in bariatric surgery.
Methods:
Aortic injuries related to stent use was the main criteria to summarize the literature by a careful assessment of PubMed/MEDLINE databases. Leak characteristics, primary endoscopic treatment, and the outcome of endoscopic complication were retrieved and categorized from each eligible article.
Results:
Thirty-five articles were selected for analysis. After abstract analysis, 22 studies were excluded, and 13 articles were reviewed in full-text version. Four articles were confirmed with aortic injury following stent use for complications after different bariatric procedures. These contained one retrospective case series and three retrospective case reports. There were 4 patients involved with complications following bariatric surgery: Roux-en-Y Gastric Bypass—3 cases and laparoscopic sleeve gastrectomy—1 case. The reported mortality of the aorto-esophageal fistula was 50%.
Conclusions:
Using stents in the treatment of leaks following bariatric surgery could be an efficient treatment, despite the poor quality of life and the stent migration. Even though it is rare, the aortic injury is a dreaded complication related to stent use and associated with high mortality rates.
Introduction
The increasing prevalence of morbid obesity and the fact that surgery is the only successful effective treatment for this condition have led to the increased number of bariatric procedures worldwide. Even if for many years Roux-en-Y Gastric Bypass (RYGBP) was considered as the “gold standard” in bariatric surgery, laparoscopic sleeve gastrectomy (LSG) has progressively become the most common bariatric procedure during the last decade, in France since 20111 and in the United States since 2013. 2 Even if LSG is considered a simple procedure, the learning curve to significantly decrease the risk of complication is flatter. The gastric leak (GL) remains the most serious early complication, mainly in case of a nonstandardized endoscopic treatment. Increased surgical experience reduces the leak rate in the recent series3,4 to 1%–2% or even less. 5
Following the diagnosis, the first line of treatment of any GL in bariatric surgery is the endoscopy. Although this aspect seems essential, only few algorithms of treatment for the endoscopic approach of the GL were published.6–8 The majority of the articles are reporting different methods of treatment (stents, pigtails, clips, or glue) without any clear indication for the choice of the device and no correlation with the leak's diameter, the presence of concomitant stenosis, the interval from surgery, and the endoscopist's experience.
In the last years, stents have evolved with commensurable improvements. However, migration rate and patient's clinical tolerance remain an important problem for stent use. Aortoesophageal fistula is an extremely rare complication of stent deployment for a complication following bariatric surgery, but considering its high percentage of mortality, it cannot be underestimated.
The aim of this systematic review was to summarize the literature data regarding the aortic injuries after stent use in bariatric surgery.
Methods
Search strategy and study selection
The search method for the identification of studies was primarily electronic. We considered also unpublished and/or non-English language studies for review inclusion. A comprehensive search of electronic databases (i.e., MEDLINE, PubMed, and ResearchGate) using broad search terms was completed. The following term combination was used: endoscopic stent, major complication, and aortic injury AND LSG or sleeve; RYGBP or bypass; and bariatric surgery. All studies reported from 2000 to December 2017 were considered. In addition, we used other resources to search for studies. The reference lists of all included studies were manually examined to identify additional potentially relevant publications. The conference abstracts, considered as “Gray literature” were analyzed for the same period.
Studies of any design that involved the treatment of leak following bariatric surgery and stent use were considered. Our initial analysis included a prescreen to identify the clearly irrelevant reports by title, abstract, and keywords of the publication. Two other independent reviewers then assessed the studies for relevance and inclusion. The studies were classified as relevant (meeting all specified inclusion criteria); possibly relevant (meeting some; but not all; inclusion criteria); and rejected (not relevant to our review). Two reviewers independently reviewed the full-text versions of all studies classified as relevant or possibly relevant. Any disagreements were resolved by repeat extraction.
Data extraction and management
Two reviewers independently extracted the data from the full version of the articles. The extracted information included details of the methods (e.g., retrospective case series), demographics (e.g., age, gender), clinical characteristics, study inclusion and exclusion criteria, number of patients excluded, number of endoscopic procedures, details of the intervention (e.g., solitary LSG, staged bariatric procedure), baseline, and postintervention outcomes.
Statistical analysis
We performed an analysis of the data from the included studies. Descriptive statistics (simple counts and mean values) were used to report the study-, patient-, and treatment level data. The number of patients enrolled was used in the calculation of the study and patient demographics. Because of the high heterogeneity among the studies and the complete lack of randomized controlled trials, a meta-analysis was not deemed appropriate.
Results
Search results
Thirty-five articles were selected for analysis using our search criteria for screening. After abstracts' analysis, 22 studies were removed, and the 13 remaining were reviewed in full-text version. Four articles were confirmed with aortic injury following stent use for complications after different bariatric procedures. These comprised one retrospective case series 9 and three retrospective case reports.10–12
Included studies
Table 1 summarizes patients' demographics and details of the endoscopic treatment with the outcome of the patients included in the four studies, identified in the current review.
Demographic Data—Summarized Results
LSG, laparoscopic sleeve gastrectomy; OR, operating room; POD, postoperative day; RYGBP, Roux-en-Y Gastric Bypass; SEMS, self-expandable metal stent.
Four cases of complications following bariatric surgery were identified: after RYGBP—3 patients and after LSG—1 patient. The reported mortality for complicated stents by aortoesophageal fistula was 50%.
In the first case, 9 a patient with perforated marginal ulcer after RYGBP was treated by endoscopy with a fully covered self-expandable metal stent (SEMS). After 1 month, the stent was removed with no sign of bleeding. Few days later, the patient returned for bacteremia with herald esophageal bleeding and an aortoesophageal fistula was diagnosed. The treatment consisted in an emergent aortic endograft placement. The postoperative recovery was long. After an admission to a long-term facility care, 1 month later the patient came back for fever secondary to acute cholecystitis and, despite percutaneous cholecystostomy, the patient died because of septic shock.
In the second case, 10 a 43-year-old female patient with a history of eroded adjustable gastric banding performed by laparotomy underwent conversion to RYGB. This revisional procedure was performed in a single stage with the band removed through a gastrostomy and simultaneously antecolic antegastric RYGB. The postoperative period was complicated by peritonitis at POD 6. During the laparoscopic exploration, the leak was identified and a closure of the defect with separate points was attempted. At POD 15, a fully covered double-bump esophageal stent (Niti-S Beta™, Taewoong Medical, Goyang-si, South Korea) was deployed for a persistent leak. The diagnosis of aortoesophageal fistula was established on postoperative day 53, when patient complained of acute hematemesis. The upper endoscopy showed a blood clot at the proximal edge of the stent, which became active bleeding spot following stent removal. The fistula was treated immediately by endovascular approach with a covered aortic stent, and a new gastrojejunostomy was performed by laparotomy. The postoperative outcome was favorable with a follow-up of 1 year.
In the third case, 11 a 55-year-old man underwent a gastric bypass complicated by gastrojejunal anastomosis stenosis and gastro-gastric fistula. The initial step of the treatment was endoscopic balloon dilation. After recurrence of the stenosis and persistent leak, a Niti-S Beta Taewoong type esophageal stent was deployed. After 2 months, he presented a massive hematemesis, with collapsing shock. After stabilization, a gastroscopy was performed allowing to remove the esophageal stent and to precise the site of bleeding in an area sitting 1 cm above the gastrojejunal anastomosis without specifying the etiology. An angioscanner diagnosed an aorta-digestive fistula and the patient was immediately treated with aortic stent. After a few days and multiple transfusions, the clinical situation considerably improved and the patient was discharged from hospital.
The fourth case 12 was a classical proximal leak at the gastroesophageal junction following LSG in a 43-year-old female. The endoscopic treatment consisted in a 12 cm fully covered SEMS for 3 weeks. After this period, the distal part of the stent was narrowed, thus the stent was replaced with another partially covered SEMS of 15 cm. Four weeks after stent deployment, a hypovolemic shock was diagnosed, presenting with hematemesis, hypotension, and tachycardia. During the emergent exploratory laparotomy, after stent removal through a distal gastrostomy, a continuous bleeding was noticed proximally at the level of the gastroesophageal junction. The stomach was completely opened and a punctiform perforation was identified in the distal esophagus, at the base of a small ulcer. The procedure was continued by a cardiovascular surgeon who performed a left thoracoabdominal incision, confirming the presence of an aortoesophageal fistula in a short segment of tense fibrosis between the distal esophagus and the aorta. The treatment consisted in a graft of the damaged area, but it was unsuccessful and the patient did not survive the procedure.
Discussion
Different anatomical particularities (gastric fundic wall is thinner, precarious vascularization of the fundus) suggest that any surgery at the level of the esophagogastric junction is associated with a more important risk of leak. In addition, the LSG is increasing the intragastric pressure; hence, the proximal area of the resected stomach is becoming more prone to a disruption of the staple line. 13 The bariatric literature and the clinical experience confirmed that the preferred location for the majority of the leaks in bariatric surgery is distally to the gastroesophageal junction. 14
After diagnosing a GL, a reasonable treatment should be implemented, consisting in the drainage of the collection, the control of all the electrolytic abnormalities associated with the sepsis management and starting an adequate alternative nutrition. After the initial stabilization of the patient, the endoscopic exploration is of outmost importance and it could represent the principal method of treatment based on a clear algorithm, including pigtails, stents, clips, or septotomies.
More than 10 years ago, the use of stents in bariatric surgery was very enthusiastic despite the initial evidence of a poor quality of life for the patients. Despite the advent of brand new longer endoscopic stents, the migration rate represents the main limitation. The literature review of this main drawback was summarized in Table 2. Different types of stent have been used, each with its advantages and inconvenients. To prevent tissue ingrowth, the fully covered stents could be preferred, but they present a higher rate of migration compared to partially covered stents. The latter ones need replacements at shorter periods to avoid difficult extraction due to the mucosal ulceration and integration of the stent into the mucosa. 22 For these reasons, the suggested initial period of 6 weeks for stent deployment was reviewed to a maximum of 3 to 4 weeks.
Literature Review with Stent Migration Rate
LSG, laparoscopic sleeve gastrectomy; RYGBP, Roux-en-Y gastric bypass.
Major aortic injuries secondary to serious mucosal erosion by SEMS were described in the literature before the expansion of bariatric surgery. Speer et al. 25 reported two major erosions for 23 patients who presented complications of leak treated with stents following cervical anastomoses after esophagectomy. In another report, Licht et al. 26 described a stent erosion into the pulmonary artery in a cohort of 31 patients who required esophageal stent deployment for leak after major oncological procedures.
Certainly, the formation of the aortoesophageal fistula implies many other factors. The main reason of erosion through the esophagus with fibrosis and formation of a fistulous tract with the aorta is represented by the radial pressure of the stent. The balance is very fragile for this radial pressure that is concomitantly preventing the stent migration. Furthermore, specific stents (Niti-S Beta stent) were designed for bariatric surgery with a double bump mechanism created to prevent migration. It is uncertain if this specific characteristic or too long deployment period (6 weeks) facilitates esophageal ulceration and consequently an aortoesophageal fistula formation. 10
Any bleeding or clot around the stent, especially in a hemodynamically unstable patient, should be cautiously diagnosed with a high index of suspicion for aortoesophageal fistula. The ideal and best workup is represented by the endovascular approach by aortic stent to control the bleeding in acute settings. This could be done in a stand-alone procedure or as an elective approach combined with a more definitive surgery to address the esophageal leak as well. 26
The treatment of leak following RYGBP is highly variable. Some authors proposed surgical reintervention with no endoscopic stenting. Bège et al. 17 reported a greater success rate, when early stenting (<30 days) was used after surgery with a more rapid healing of the leak and a diminished number of endoscopic procedures.
Recently, Moon et al. 27 reported the stent experience of 87 patients with complications following different bariatric procedures (33 RYGB, 50 LSG, 2 DS, and 2 vertical banded gastroplasty) in four bariatric centers for a period of 4 years (2013–2017). For 17 patients (19.5%), repositioning or stent replacement was needed because of the initial stent migration. After definitive stent removal, for 13.8% of patients, a stenosis was diagnosed and further treated with repetitive endoscopic dilatation.
Thompson and colleagues 18 have reported a comprehensive review regarding the use of stents in the treatment of bariatric surgery leaks. At that time, no case of aortic injury or aortoesophageal fistula after stent use was reported. Our current review identified 4 cases of such dreaded complication of stent use in bariatric surgery, and the statement of “no stent associated mortality” must be revised. Surprisingly, 3 out 4 cases were following RYGBP. Probably many of the complications following LSG were also case reports, and probably, they were unreported.
Conclusion
The stents are effective tools for the management of leaks after foregut bariatric surgery. The biggest challenge with this approach is stent migration and poor quality of life. Caution is warranted due to the risk of fatal complications such as aortoesophageal fistula formation. Aortic injury is a rare but serious complication after stent use. Although it is rare, bariatric surgeons and endoscopists should be aware of this life-threatening complication.
Footnotes
Disclosure Statement
No competing financial interests exist.
Funding Information
No funding was received for this manuscript.