Completion Gastrectomy with Esophagojejunostomy for Management of Complications of Benign Foregut Surgery

Introduction

Foregut surgery comprises a spectrum of operations from peptic ulcer disease (PUD) to antireflux fundoplications, and with the worldwide epidemic of obesity, an increasing number of bariatric operations. In experienced centers, primary foregut surgery can be completed open or laparoscopically with low morbidity. Reported overall complication rates range from 5.9% to 10.6% in bariatric surgery and from 2.1% to 8.3% in antireflux surgery.^1–4 However, the increased number of patients with prior foregut operations has resulted in an increase in revisional foregut surgery. The incidence of nonbariatric revisional foregut surgery is difficult to estimate. The American Society for Metabolic and Bariatric Surgery (ASMBS) estimated that >9000 bariatric revision surgeries were performed in 2011. This has increased exponentially to >26,000 estimated in 2015, and is projected to further increase. ⁵

The most common indications for revisional foregut surgery are undesirable results of the primary surgeries, such as failed weight loss after bariatric surgery or complications from the initial surgeries. However, the distorted anatomy and tissue planes obscured by adhesions or inflammation create challenging operating conditions and contribute to complication rates as high as 39.3% in some series of revisional foregut surgery.^6–8 Completion gastrectomy with an esophagojejunostomy (EJ) in this setting has been described as a last resort for proximal chronic leak after sleeve gastrectomy. In some cases, dissection around the hiatus compromises blood flow to proximal stomach, making EJ unavoidable.^9–12 Published evidence in the modern literature is relatively scarce and the historical literature often includes total gastrectomy to treat malignancies, or to treat rare diseases such as Zollinger–Ellison Syndrome. ¹³ Owing to the infrequency of this event, we sought to describe our experience. The primary objective of this study is to report our tertiary referral experience with patients undergoing total or completion gastrectomy with Roux-en-Y EJ reconstruction, indications for surgery, and 30-day and 6-month outcomes to assess the feasibility and morbidity of this uncommon operation after benign foregut surgery.

Methods

After approval by Institutional Review Board, all patients who underwent total, proximal, or completion gastrectomy with Roux-en-Y EJ reconstruction for complications of benign foregut surgery between January 2006 and December 2015 were retrospectively identified from the electronic medical record. Patients who underwent operations for suspected or for confirmed malignancies were excluded from the study. The type of initial surgery, complications, and symptomatology that lead to revisional operations, previous surgical and endoscopic interventions before EJ, and preoperative nutrition status were collected. In addition, 30-day perioperative complications and 6-month postoperative outcomes using Clavien–Dindo classification, postoperative need for nutritional support through either enteral feeding or total parenteral nutrition (TPN), and 30-day readmission were assessed. Complications classified with Clavien–Dindo ≥ III were considered major postoperative complications. ¹⁴

Operative technique

As this study includes various revisional surgeries, it is challenging to detail all operative techniques. The open and laparoscopic techniques typically use similar instrumentation and follow similar steps, which are briefly outlined hereunder.

For the laparoscopic technique, access to the peritoneal cavity is gained and pneumoperitoneum established. Three 5 mm, one 12 mm, and one 15 mm trocars are typically used, as well as a liver retractor in a configuration depicted in Figure 1. For the open approach, a left-sided thoracoabdominal incision or vertical midline laparotomy is used depending on surgeon preference. Typically, adhesions between the anterior surface of the stomach or gastric pouch and the left lobe of the liver must be fully divided to allow anterior retraction of the liver and placement of a liver retractor. Significant adhesions are typically present around the diaphragmatic hiatus, and these are divided with the goal to observe the right and left crus of the diaphragm. Once the crura are identified, the esophagus is mobilized circumferentially. Securing a Penrose drain around the distal esophagus at this point aids in retraction. It is imperative to dissect the esophagus within the posterior mediastinum to assure sufficient intraabdominal length of the esophagus in order to fashion EJ. An intraoperative upper endoscopy helps to identify healthy, uninflamed tissue of the distal esophagus and/or to identify whether internal openings of chronic fistulae are present. This is particularly important when endoscopic clips have been previously placed, so as to avoid firing a stapler across them. One or two firings of linear stapler (3.5 mm staple heights) are used to transect the distal esophagus. The stomach or gastric pouch is also divided distally in an area clear of inflammation, with a linear stapler. Reconstruction then takes place, using a 75–150 cm Roux limb, based on surgeon preference and in consideration of the patient's body mass index (BMI). The decision to place the Roux limb in a retrocolic or antecolic position typically depends on the reach of the Roux limb, with the final position chosen to minimize tension on the anastomosis. The EJ anastomosis is most commonly made with a 25 mm circular end-to-end anastomosis (EEA) stapler (3.5 mm staple height). To complete this, an anvil is introduced per os and passed through a small esophagotomy. The EEA stapler is introduced through the left upper quadrant port site in the laparoscopic approach, or directly into the abdomen in the open approach. The terminal staple line of the Roux limb is opened, and the EEA stapler is passed into the lumen, mated to the anvil and fired, creating an end to side, functional end-to-end EJ anastomosis. A single firing of linear stapler (3.5 mm staple heights) is used to close the candy cane. Also included in this series are hand sewn and linear stapled anastomoses. Medial and lateral sutures are placed from the Roux limb to the right and left crura to decrease the tension at the anastomosis. The jejuno-jejunal anastomosis was made in a side-to-side functional end-to-side configuration using linear staplers or a hand-sewn technique, with a 50 cm biliopancreatic limb. Feeding jejunostomy is placed in the Roux limb, at the surgeon's discretion. At least one closed suction drain is placed near the anastomosis.

OPEN IN VIEWER

FIG. 1.

Laparoscopic port placement. Three 5 mm, one 12 mm, one 15 mm, and a liver retractor are typically used for the laparoscopic approach.

Results

During the study period, 864 gastric operations were performed for both benign disease and malignancies. Operations for benign disease include bariatric surgery, PUD, benign tumor, and antireflux operation. Twenty-three patients underwent completion gastrectomy with Roux-en-Y EJ, including 13 patients who underwent laparoscopic EJ (LEJ) and 10 who underwent OEJ. One case that was started laparoscopically was converted to open case due to severe adhesions, and was categorized as open for the remainder of the analyses due to the physiologic insult of an open operation. The index operations included 12 antireflux, 9 bariatric, and 2 PUD surgeries. Patient demographics are given in Table 1. The group had a mean age of 55 ± 2.8 years and mean BMI of 30 ± 1.5 at the time of their EJ operation. Indications for revisional surgery are listed in Table 2. Most common reason for revision was failed antireflux operation followed by anastomotic complications, such as anastomotic leak or stricture from the index operation. The patients, who required EJ for failed antireflux operations in the presented study, included those who developed postoperative complications, or recurrent symptoms such as severe dysphagia and failure to thrive. They were unable to undergo the revision of antireflux operation nor the conversion to Roux-en-Y gastric bypass due to severe inflammation or adhesions that made dissection extremely challenging secondary to the prior multiple fundoplications (range 1–3). Seventy-eight percent of patients had prior surgical or endoscopic interventions, including 9 patients with various endoluminal treatments, 16 patients with at least one prior revisional surgery, and 7 patients who underwent both endoluminal and surgical revision before EJ. Endoluminal treatment included pneumatic dilatation, endoluminal stent placement, endoscopic suturing, and over-the-scope clip placement, and in some cases multiple interventions were performed in a single session. Mean number of interventions before EJ was 3 ± 0.8 times (range: 0–15) and median interval from the initial operation to EJ was 25 months (IQR 9–87) (Table 1). Approximately 40% of patients required preoperative enteral or parenteral nutrition therapy, secondary to malnutrition. Preoperative nutritional intervention was determined based on a multidisciplinary assessment involving a nutritional support team and the surgical team. This was based on a combination of factors, including objective weight loss, evidence of sarcopenia, serum laboratory testing, and a consideration of comorbid conditions.

As shown in Table 3, comparison between LEJ and OEJ showed similar operative time, estimated blood loss, and overall complication rates. LEJ was associated with a shorter length of stay (LOS) (P < .001), fewer postoperative ICU days (P = .002), lower 6-month complication rate (P = .007), and less likely readmission (P = .02). The surgical approach at the index operation had no bearing on the surgical approach used for EJ (P = .21). Laparoscopic or open approach was based on surgeon preference, experience, but tended toward laparoscopic approach later in the series. An antecolic position of the Roux limb was more frequently chosen in LEJ, and retrocolic Roux limb position was more common in OEJ. The majority of EJ anastomoses were made with a circular EEA stapler in both groups (82.6%). Postoperative nutrition therapy, defined as either enteral tube feeds or TPN, was required after all OEJ and 46% of LEJ (n = 6/13), majority of whom received enteral nutrition. Postoperative nutritional intervention was initiated based on the surgeons' preference, patients' postoperative oral intake, and the recommendation from dietitians.

The 30-day perioperative complication rate was 30% (n = 7), with 17% (n = 4) classified as major (Clavien–Dindo ≥ III) (Table 3). Although there was no 30-day perioperative mortality, there was one nonprocedure-related mortality within 6-month postoperative period. Postoperative complications within 30 days and 6 months are given on Table 4. Anastomotic stricture requiring endoscopic dilation was the most common complication in this series, and all of them occurred beyond the immediate postoperative period. Long-term major complications that require further interventions were more common in OEJ (Table 4).

Discussion

In this study, we performed 23 total and completion gastrectomy with Roux-en-Y EJ reconstruction for complications of benign foregut surgery. This is the largest case series of EJ for benign foregut surgery complications.^9–12 Total gastrectomy with Roux-en-Y EJ reconstruction is often considered for foregut malignancies, such as cancer involving the gastroesophageal junction or proximal stomach; it is rarely performed for benign disease given its complexity and morbidity.^15–17 Moreover, many patients do not tolerate a lack of a stomach well, and an EJ anastomosis may be associated with dumping syndrome, reflux, or weight loss. ¹⁸ However, completion gastrectomy with EJ has recently been reported as a definitive treatment for complicated foregut surgery, especially chronic sleeve leak, with promising results in small series.^9–12 The objective of this case series was to report feasibility and safety of completion gastrectomy with EJ reconstruction for benign foregut surgery complications as a salvage operation, potentially avoiding esophagectomy.

Complication of prior foregut operation is the most common indication for revision.^6–8,19 In many cases, these patients have had limited nutrition, and may have chronic infections or fistulae. Each revisional case is unique, and each merits a carefully considered operative plan. Preoperative assessment is of utmost importance before proceeding to this complex and challenging operation, usually in conjunction with a multidisciplinary team that includes an experienced dietitian, internist, radiologist, and surgeon.^6,7 The majority of the patients in this series underwent an average of three interventions over 25 months before EJ reconstruction, and 70% of them had at least one previous surgical intervention. Nutritional status should be evaluated before planned intervention, and preoperative supplementation with enteral or parenteral nutrition may be necessary. For example, this series includes 9 patients who received preoperative nutritional therapy. Mean preoperative serum albumin and serum prealbumin in all cohorts were 4.0 ± 0.1 g/dL and 16.9 ± 1.5 mg/dL, respectively. Active infection also must be managed with proper source control, drainage, and appropriate antibiotic coverage before attempting EJ reconstruction. This is particularly true in patients with chronic anastomotic leak before EJ.^9–12

Despite preoperative assessment, major perioperative complications in this series were approximately 17% without 30-day mortality. Despite this challenging operation in the reoperative fields, this complication rate seems compatible with the current literature.^{6–8,20–22} Anastomotic leak occurred in 2 patients (8.6%), both of whom required reoperation for surgical drain placement. One patient with an anastomotic leak underwent esophageal reconstruction with a colonic interposition a year later. Anastomotic stricture requiring pneumatic dilatation (n = 5, 21.7%) was the most common postoperative complication, requiring a median of one endoscopic pneumatic dilatation (range 0–2) during the follow-up period. The majority of anastomotic strictures responded to a single pneumatic dilatation without further interventions. We acknowledge that the stricture rate in our study is higher than stricture rates estimated between 2.2% to 11.9% reported in series of patients undergoing gastrectomy for malignancies in Korea and Japan.^23,24 First, this series has a longer follow-up period, whereas most other publications focus on perioperative or early postoperative complications. Second, all of the cases in this series are undertaken in the setting of at least one prior foregut operation. As such, the EJ anastomosis is placed in the region of dense inflammatory tissue that exists at the hiatus, which cannot be ignored as a contributing factor. Lastly, stricture rates are also dependent on the technique used to create the EJ anastomosis, and thus direct comparison may not be possible. The one mortality in the cohort was nonprocedure-related cardiac complication during the recovery in the rehabilitation facility after discharge from the hospital.

Postoperatively, 70% of the patients (n = 16) required nutritional support with either parenteral or enteral nutrition, which included the 9 patients given preoperative supplementation. Establishing enteral feeding access at the time of EJ operation, such as jejunostomy tube or remnant gastrostomy tube, is a reasonable option to provide postoperative supplemental nutrition. ²⁵ If needed, short-term parenteral nutrition has been shown to be beneficial for complications of bariatric surgery, and extrapolation can be made to this population.^26,27 As Van Gossum et al. recently reported, 77 out of 2800 home PN patients had history of bariatric surgery and 46 patients successfully weaned off their PN in 7–10 months. ²⁵ Most patients in our series (75%) were able to wean from nutritional support in an average of 4 months (range 1–14 months); however, 6 patients continued to require enteral nutrition support due to insufficient oral intake at 6 months after the EJ.

When compared with an open approach, LEJ was associated with shorter LOS, fewer ICU days, less long-term morbidity, and less 30-day readmission rate. Typically, the OEJ group included patients with thoracoabdominal incision who often required ICU management postoperatively, whereas LEJ patients were usually managed on the regular nursing floor. Splinting from chest wall pain, the presence of chest tubes, and a longer cumulative incision length in the OEJ group were likely contributing factors to longer ICU and hospital stays. Laparoscopic approach likely enabled earlier mobilization, less postoperative pain with a subsequently lower opioid requirement, and earlier return of bowel function.

The ASMBS currently recommends revisional bariatric surgery to be performed in high-volume facilities due to the complexity and higher associated morbidity. ⁷ Although this recommendation targets revisional bariatric surgery, patients needing revision of prior antireflux and PUD operations should likely follow the same recommendation. Where expertise exists, both the results of this study and other recent literature suggest that the minimally invasive approach is preferred.^9–12 Since the first LEJ for chronic leak after sleeve gastrectomy was performed in 2012, LEJ was more commonly performed later in this case series, even despite prior open surgical interventions, as given in Table 3 (approximately 40% of LEJ patients had previous open surgery).

Historical convention dictates that the Roux limb should be brought retrocolic when making an EJ anastomosis, to create additional length and decrease anastomotic tension. However, as the laparoscopic approach for total gastrectomy and Roux-en-Y gastric bypass has become more common, antecolic Roux reconstruction has been more accepted due to simplicity without compromising the outcome. ²⁸ Hence, antecolic approach is more often elected in the LEJ group.

Limitations of our study include the small sample size, which limits comparisons between laparoscopic and open approaches. Although this cohort is also heterogeneous in terms of the index operations, complications, and indications for revision, this is also a fair and accurate representation of this patient population. A variety of techniques were used to make the EJ anastomosis, as well as varying positions of the Roux limb, and comparisons are limited by small sample sizes. These data were accrued over a 10-year period; however, a great deal of variability existed in terms of follow-up. Because operations were performed at a tertiary referral center, several patients were completely lost to follow-up beyond 6 months after their operations, which introduces a selection bias.

In conclusion, our series demonstrates that Roux-en-Y EJ is a reasonable option for reoperative foregut surgery, and can be completed with either open or laparoscopic approach. The laparoscopic approach appears to be associated with decreased LOS, ICU utilization, and readmissions.