Innovative Use of Atrial Septal Defect Occluders for Gastric Leak Management After Sleeve Gastrectomy: A Case Report

Introduction

A sleeve gastrectomy is one of the most performed bariatric procedures worldwide [1], and it has become a very frequent procedure in bariatric surgery due to its simplicity and efficacy compared with other procedures [2,3]. Unfortunately, the postoperative leak rate can vary between 1% and 3% for primary procedures [4], and more than 10% for revision procedures [5,6].

The management of gastric leaks is complex and not yet standardized. Management is particularly challenging due to several factors including the high-pressure nature of the area, which hinders spontaneous closure; surrounding inflammation, which makes surgical revision more difficult; and malnutrition resulting from reduced oral intake, which further delays tissue healing. Several different treatment strategies have been described and include both surgical and endoscopic approaches [7]. The first consists of a laparoscopic revision that includes resuturing and drainage. The second includes the endoscopic delivery of different types of stents, closure devices, internal drainages, and suturing [7,8]. While these traditional methods have established roles in clinical practice, they are frequently limited by technical difficulties, high migration rates, or invasiveness.

We present a case of gastric leak after sleeve gastrectomy that was initially treated with a self-expandable, covered mega stent but was further complicated by its migration. We describe our management of this case of persistent leak after sleeve gastrectomy due to failure of initial management and consistent refusal of surgery by the patient. To do this, we employed a novel endoscopic technique using an atrial septal defect (ASD)-occluding device (ASD occluder). By documenting this approach, we aim to expand the therapeutic options available to clinicians managing this challenging complication, particularly in cases refractory to standard therapies.

Case Report

A 20-year-old Saudi woman underwent a laparoscopic sleeve gastrectomy operation because of morbid obesity, in another hospital (not the hospital of the authors of this report). A few days after the operation, the patient experienced abdominal pain, tenderness, and vomiting. She returned to the hospital where the surgery had been conducted, where they managed a gastric leak with the placement of a mega stent and drainage of the small collection she had developed earlier.

Ten days later, the patient presented to the emergency department at our hospital with a 1-day history of abdominal pain, primarily in the left upper quadrant. The pain was severe, constant, and progressive, was not aggravated or relieved by any specific factors, and radiated to the left shoulder. She also reported recurrent vomiting and a subjective fever.

Upon examination, the patient appeared dehydrated. Her vital signs showed a rapid pulse rate of 113 bpm, normal blood pressure, and oxygen saturation of 93% on room air. A local abdominal exam revealed a soft and lax abdomen with mild tenderness in the left upper quadrant.

Her initial blood work revealed a white blood cell (WBC) count of 17.9 (elevated), hemoglobin of 12.3, platelet count of 437, international normalized ratio (INR) of 1.55, and C-reactive protein of 278 (elevated). The renal function tests and lactate level were normal.

A computed tomography (CT) scan with contrast showed gastric stenting in the residual stomach without evidence of a definite contrast leak. A large pocket of air was noted in the lesser sac, likely related to the prior leakage. Additionally, there was an 11.7×10.6×3.2 cm sub-capsular splenic collection causing left basal lung consolidation.

On the same day, ultrasound-guided drainage of the subcapsular collection around the spleen was performed, yielding approximately 70 ml of pus on the spot. A drain was subsequently placed and left in position.

A Gastrografin study was performed, which showed no significant obstruction. However, a mild delayed passage was noted at the level of the esophageal gastric stent. Definitive extraluminal leakage of contrast media was observed. Based on these findings, the patient was not allowed to eat by mouth, and started on total parenteral nutrition (TPN).

Later, an initial upper endoscopy was done to evaluate the stent position. The endoscopy showed esophageal leakage. There was a fully covered metallic mega stent in place, and at 27 cm, a visible esophageal leak through the stent was noted. The distal end of the stent was located at the duodenal bulb. Consequently, the stent was pulled proximally by 5 cm, resulting in the new proximal end being positioned at 20 cm. After 1 week, a Gastrografin study revealed contrast leakage through the middle part of the stent.

A second upper endoscopy was performed, revealing an esophageal stent in place with its proximal end located 25 cm from the incisors under suction. Additionally, fluids were observed through the stent at 27–28 cm from the incisors, underneath the stent, consistent with findings from the previous endoscopy. The stent was removed using rat-tooth forceps.

Examination of the surgical site revealed a 1 cm defect at the gastroesophageal junction, 30 cm from the incisors. Argon plasma coagulation was applied to the edges for marking, and an Ovesco clip was placed. The defect was successfully closed, as documented in the accompanying images. Pressure ulcers caused by the distal end of the stent were noted in the duodenal bulb, but no bleeding was observed.

A third upper endoscopy was performed, revealing that the Ovesco clip over the gastric leak had loosened and was subsequently removed. The leak had a diameter of approximately 7 mm, and the opening behind it appeared smaller than before and clean.

The patient underwent a fourth endoscopic procedure, which revealed a leak at the site of the previous defect, measuring approximately 7 cm in diameter. This time, we used a novel ASD device (Amplatzer™; size 13; Reference number: 9-ASD-013) to close the leak, and the procedure was successful (Video 1). The ASD device was detached from its delivery system. Using a therapeutic gastroscope, we navigated a biliary catheter through the working channel to recapture the device with pediatric biopsy forceps. Once properly positioned, we advanced the device into the fistula and deployed the distal flange. The proximal flange was then successfully opened on the luminal side of the fistula, completing the placement.

Three days later, we arranged for CT with intravenous and oral contrast and it showed no evidence of leaking. On the same day, a Gastrografin study also reported no evidence of a leak.

A week later, we performed an upper endoscopy and found that the esophagus was normal. The stomach showed the ASD closure in place with minimal tissue growth, and we did not observe any leak. The duodenum was also normal. We started the patient on sips of water and gradually advanced her diet to clear liquids and then to full liquids as tolerated, while we monitored for any clinical signs of a leak. The patient dramatically improved after the ASD device procedure and was advanced to a full diet, and was finally discharged home with medication, multivitamins, and clinical follow-up.

The patient was later seen in the clinic and was in good condition. We arranged for a follow-up endoscopy to be performed on the same day as the appointment. On the 27th of October 2022, an upper endoscopy was performed to confirm that the ASD device was in place. The endoscopy showed the ASD device in position, with minimal tissue growth around it (Figure 1).

No adverse effects or complications were observed during or after the procedure. The patient remained asymptomatic for 4 weeks after the first presentation.

One month later, the patient presented to the emergency department again with left shoulder pain. The CT scan showed a larger-size collection compared with the previously seen gastrosplenic collection, as well as free air and contrast extravasation suggestive of persistent leakage (Figure 2).

A Gastrografin study reported extraluminal leakage of contrast media at the level of the fundus of the stomach extending toward the left subphrenic space. A peripherally inserted central catheter (PICC) line was inserted, the patient was started on TPN, and drainage for the collection was done.

Esophagogastroduodenoscopy (EGD) was done and the ASD was easily removed. The tissue underneath appeared healthy, with good granulation tissue, and a guide wire was advanced under fluoroscopy guidance to the jejunum. An 8 French nasojejunal tube was inserted under fluoroscopy and positioned correctly. Another EGD was repeated and an esophageal leak was seen at 32 cm from the incisors, no pus was seen, and the nasojejunal tube was removed. A mega stent of 18 cm was inserted under fluoroscopy and guidewire guidance, and the proximal end of the stent was placed at 27 cm from the incisors. The new CT showed focal oral contrast noted within the subhepatic collection, highly suggestive of a current leak. The gastric stent was in place. There was a mild decrease in the size of the subhepatic collection.

The patient underwent another EGD procedure, in which the stent was seen at 28 cm from the incisors and was removed by a rat-tooth tool. Ulcerations were seen at the proximal and distal end of the stent. After removal, the gastroesophageal junction was seen at 34 cm from the incisors and the leak was seen just above the gastroesophageal junction, with pus discharge.

After the patient exhausted all nonoperative management options, a family meeting was conducted and the family finally agreed to the Roux-en-Y bypass. After multiple follow-up visits, the patient improved and no evidence of a leak was detected.

Discussion

According to recent literature, leakage is a significant complication following sleeve gastrectomy. It occurs in 1.85–2.7% of patients and has a mortality rate ranging from 0.4–3.7% [9]. Gastric leaks may be caused by multiple factors, including technical, mechanical, and ischemic causes. Stapler misfiring, defects in sleeve stapling and suturing, devascularization of the gastric fundus at the angle of His, gastric stenosis, and kinking are all factors contributing to its genesis [9].

A gastric leak may manifest as acute (occurring within 7 days), early (occurring within 1 to 6 weeks), late (occurring within 6 to 12 weeks), or chronic (occurring beyond 12 weeks) depending on when it presents. The severity and clinical appearance are directly linked to the nature of the staple-line leak [10]. The cornerstone methods for treating leaks following sleeve gastrectomy are surgical lavage and drainage or percutaneous drainage of any intra-abdominal collection and installation of a covered stent. Other less invasive methods, like pig-tail drainages, vacuum systems, and over-the-scope clips, have been introduced recently to treat staple-line leaks [11,12].

The ASD occluder is a circular disk composed of nitinol wire mesh that self-expands and resembles a double umbrella. A brief connecting mesh tube, which serves to stent the defect, connects the 2 disks. In general, treatments for congenital cardiac abnormalities are extensive. Because there is a wide range of sizes available for the devices, the defect can be effectively closed right away by matching the device to the exact length and diameter of the lesion. An essential advantage of the ASD occluder is its ease of deployment, retrieval, and repositioning before detachment, which ensures precise placement and effective coverage of the fistula. While it is new, the application of an ASD occluder to seal a gastrocutaneous fistula as detailed below was documented in the setting of a bronchopleural fistula via an endobronchial approach, and gastro-colonic fistula and biliodigestive fistula between the bulbous and common bile duct via an endoscopic approach [13,14].

In our case, the decision to use the Amplatzer septal occluder device was due to failed initial management and patient refusal to perform surgery. In a report by Odemis et al, the decision was made after 2 failed endoscopic repair attempts within a 5-day interval with the Over-the-Scope Clip application (nontraumatic type). Similarly, the study concluded that the device was successful in the management of gastrocutaneous fistula [15]. The technical adaptation required to deploy this cardiac device in the gastric environment, including modified sizing and deployment techniques, provides valuable procedural insights for interventionalists facing similarly challenging cases.

Advanced endoscopic procedures for localization and guidewire cannulation of the fistula could be used to place the device. Under endoscopic and fluoroscopic guidance, we were able to achieve accurate positioning and orientation in our instance. The ASD occluder was then able to be moved, and once the occluder spontaneously retracted into the stomach wall and expanded in this position, the fistula closed entirely. In addition, the patient tolerated the procedure well and experienced no difficulties or adverse effects. The septal occluder may serve as a definitive, single-step solution for post-sleeve gastrectomy leaks rather than merely a temporary bridge to surgery, potentially sparing patients from the morbidity of multiple interventions and offering a permanent closure option in cases where conventional approaches have failed.

A systematic review was conducted by De Moura et al to assess the success of cardiac septal occluders (CSOs) in the management of fistulas. The study included 22 cases of different types of fistulas: esophagopleural, duodenocutaneous, tracheoesophageal, and those occurring after sleeve gastrectomy. They concluded that CSOs have a 100% technical success rate, a 77.27% clinical success rate, and a 22.72% adverse event rate, with no deaths related to their off-label use. Moreover, it was found that there were no correlations between patient age, fistula size, fistula duration, prior treatment, and adjunctive therapy with successful fistula closure or adverse events [16]. Furthermore, other recent studies in the literature showed that use of a CSO was successful in the management of bronchopleural fistula [17,18].

The application of ASD occluders for managing gastric leaks remains in its early stages, with our case highlighting both the potential and the constraints of this approach. Despite initial success in sealing the gastric leak, our case revealed significant limitations as the ASD occluder failed to provide a long-term solution, ultimately necessitating Roux-en-Y gastric bypass after recurrent symptoms and persistent leakage. A major technical limitation is the procedure’s complexity, requiring specialized skills in endoscopic and fluoroscopic guidance, which may restrict widespread adoption. Furthermore, the risk of device migration or failure, as demonstrated in our case, poses concerns for patient safety.

Future studies should incorporate a control group or comparative analysis with established treatments like surgical repair or endoscopic stenting to objectively assess the ASD occluder’s effectiveness. We recommend implementing standardized long-term follow-up protocols of at least 12 months, including regular clinical assessments, imaging studies, and patient-reported outcomes to evaluate sustained leak closure and identify delayed complications. Additionally, developing detailed procedural guidelines with step-by-step descriptions of endoscopic and fluoroscopic techniques would improve reproducibility, while risk stratification systems could help identify ideal candidates for this intervention versus those requiring alternative approaches.

Conclusions

We demonstrated a brand-new endoscopic method that uses an ASD occluder to manage leaks after gastric sleeve surgery. Although the device did not show complete effectiveness in managing our case, we anticipate that this device will be used in clinical trials in the future to treat many inherited, congenital gastrointestinal tract abnormalities, or complications from other surgeries, as many studies in the literature have shown the reliability and effectiveness of the device in managing different types of fistulas.