Laparoscopy-Assisted Transjejunal ERCP in a Patient with Roux-en-Y Reconstruction Following Partial Gastrectomy *

Background

Roux-en-Y reconstruction is routinely employed after resections of hepatobiliary or gastroduodenal segments. In recent years, Roux-en-Y gastric bypass (RYGB) has become the standard for bariatric surgery, accounting for more than two thirds of all such operations performed annually in the United States. ¹ Consequently, endoscopists are more often faced with the need to carry out endoscopic examination and therapeutic procedures, such as endoscopic retrograde cholangiopancreatography (ERCP), in patients with Roux-en-Y anatomy. Although several techniques have been described for nonsurgical access to the duodenum in those patients, advancement of the duodenal endoscope through the anatomic route is invariably challenging and frequently unsuccessful. ²

Laparoscopy-assisted ERCP has been described as an option after gastric bypass with Roux-en-Y reconstruction. In those patients, the distal stomach is the usual point of entry after gastrostomy is performed laparoscopically.^3–9 However, the transgastric approach is not an option in patients who have undergone Roux-en-Y reconstruction after partial gastrectomy for peptic ulcer disease. In those patients, a long Roux limb (60–75 cm) of the small bowel is created to prevent bile reflux into the stomach. The length of small bowel to be traversed usually exceeds the length of the duodenal endoscope, excluding advancement through the anatomic route as a viable option. In this article, we report the case of a patient who underwent successful laparoscopy-assisted ERCP through an enterotomy, rather than the usual gastrostomy, approach. In addition to allowing for performance of the endoscopic intervention, laparoscopic examination made it possible to diagnose and treat an internal hernia, which occurred as a complication of the remote partial gastrectomy with Roux-en-Y reconstruction.

Case Report

An 18-year-old Caucasian female with a history of recurrent abdominal pain was admitted to the Gastroenterology Service with nausea, vomiting, and a new episode of abdominal pain. Her symptoms were intermittent since onset 3 years prior but became significantly more severe within the past 3 months. The patient had undergone multiple abdominal operations, including Nissen fundoplication at age 12 for congenital hiatal hernia, laparoscopic cholecystectomy at age 15 for intermittent abdominal pain, and laparoscopic partial gastrectomy with Roux-en-Y reconstruction at age 16 for presumed bile reflux gastropathy. All operations were performed at the referring institution prior to presentation at our institution. Laboratory tests obtained at different periods of time while the patient was symptomatic repeatedly demonstrated elevated liver chemistries (LFTs) in a cholestatic pattern. The LFTs normalized after the resolution of symptoms. The pancreatic enzymes were always normal. Repeated computed tomography (CT) and magnetic resonance evaluations showed normal biliopancreatic anatomy, failing to demonstrate any abnormality to account for her symptoms. Based on the transient elevation of liver tests with abdominal pain, the diagnosis of sphincter of Oddi dysfunction (SOD) was considered.

At our institution, ERCP was attempted. The apparent Roux limb was entered but only for a short distance. A percutaneous cholangiopancreatogram (PTC) was obtained and was found to be unremarkable. The PTC catheter was left in place, crossing the ampulla into the duodenum. Another attempt was made to reach the papilla, but the same situation resulted.

The patient agreed to undergo a laparoscopy-assisted ERCP, which was undertaken 3 days latter. After the pneumoperitoneum was created, the patient experienced a drop in blood pressure and decrease in oxygen saturation. A right pneumothorax was diagnosed and promptly treated with the insertion of a small-bore chest tube. It was formed by the tracking of air from the peritoneal cavity along the recently placed PTC catheter, which traversed the right diaphragm. Her blood pressure and oxygen saturation immediately normalized, and the procedure resumed.

During the laparoscopic examination, an adhesion was noted to have formed from the cut edge of the mesentery of the Roux limb down to the mesentery of the biliopancreatic limb. This led to the formation of an internal hernia with incarceration of all but the very distal portion of the terminal ileum. A subperitoneal hemorrhage of the mesentery was noted, but not a mesenteric infarction. The bowel was reduced and the adhesion forming the hernia was divided. Subsequently, the biliopancreatic limb was identified proximal to the jejunojejunostomy. A purse string of a 2-0 absorbable suture was placed, and an enterostomy was created. The 5-mm left-upper quadrant port was replaced with a 15-mm port to allow for insertion of the duodenoscope. The ampulla was endoscopically identified and biliary sphyncterotomy was performed with a needle knife over the PTC catheter, yielding excellent drainage of bile. The scope was withdrawn from the enterotomy. The bowel was then closed with absorbable sutures transversely in two layers to minimize the risk of stricturing.

Two days postoperatively, the patient complained of abdominal pain and was found to have a low-grade fever. Repeat CT scan examination revealed two small fluid collections in the dome of the liver, thought to be consistent with small liver abscess. In addition, a small branch of the right portal vein was clotted. Intravenous antibiotics and enoxiparin were started. Her symptoms resolved, and she was able to tolerate a regular diet without difficulties. She was discharge home to complete a course of oral antibiotics and oral anticoagulation. She was seen in follow-up 2 weeks after discharge, where a repeat CT showed resolution of the liver abscess and stability of the small portal clot. The patient has remained completely asymptomatic with follow-up to 3 months.

Discussion

Roux-en-Y reconstruction is routinely performed during bariatric surgery and occasionally after gastrectomy. On occasion, endoscopists are faced with the need to carry out an endoscopic examination and interventions in patients with altered anatomy due to surgery. When ERCP is indicated, assess to the duodenum through the anatomic route is frequently challenging. Wright et al. have described techniques aimed at facilitating nonsurgical access to the duodenum in patients after Roux-en-Y reconstruction, which allowed for successful ERCP in 55% of their patients. ² The large number of failed procedures, even in very experienced hands, has motivated endoscopists to explore other alternatives.

Enteroscopes and colonoscopes have been successfully used to perform ERCP in patients after bariatric surgery.^10–13 Unfortunately, only selected reports are available describing successful cannulation of an intact papilla.^11–14 To our knowledge, there is no formal analysis of success rates in reaching the duodenum or cannulating the papilla. In addition, case reports underline that such attempts are not devoid of complications. ¹⁴ Other creative approaches to access the duodenum in patients with Roux-en-Y anatomy have been reported. Insertion of the duodenoscope through mature gastrostomy or feeding jejunostomy tracts is possible when feeding tubes are in place, but not practical when they are not.^17–19

More recently, the use of double- (DBE) and single-balloon enteroscopes (SBE) has been reported in patients with surgically altered anatomy.^19–27 In patients with Roux-en-Y anatomy, DBE allowed access into the Roux limb in approximately 90% of patients. ERCP was successfully performed in 77% (20/26).^20–27 Koornstra acknowledges that cannulation after hepatojejunostomy is significantly easier than of a native papilla when the forward viewing DBE is used. That was the case with the great majority of reported patients, thus the high cannulation success rate. In addition, he calls attention to the lack of therapeutic instruments long enough for use with the DBE. ²⁰

Following Peters et al.'s first description of the technique in 2002, five other reports were published describing laparoscopy-assisted transgastric endoscopy, all undertaken in patients post-RYGB.^3–5 Laparoscopy-assisted transgastrostomy ERCP was attempted in a total of 17 patients. One procedure was not successful due to an impacted stone in the ampulla, which could not be retrieved endoscopically. ⁶ Conversion from laparoscopic to open procedures due to adhesions was necessary in 2 patients who had undergone open RYGB. After creation of the surgical gastrostomy, ERCP was performed with biliary sphyncterotomy and stone extraction. ⁸

Multignani et al. reported the only case of laparoscopy-assisted ERCP in which the duodenoscope was advanced through an enterotomy. Their patient, a 63-year-old female, had undergone RYGB 4 years earlier. She developed obstructive jaundice due to an impacted stone in the common bile duct. The duodenoscope was advanced through an enterostomy, which was created laparoscopically, distal to the ligament of Treitz. A guidewire was laparoscopically advanced into the cystic duct, and bile duct cannulation was achieved by using the rendezvous technique. Endoscopic sphincterotomy and extraction of stones were then successfully performed. ⁹

To our knowledge, our report is the first to describe a patient undergoing laparoscopy-assisted transjejunal ERCP with Roux-en-Y reconstruction for reasons other than RYGB. In patients such as ours, who have undergone partial gastrectomy with Roux-en-Y reconstruction to avoid bile reflux, the transgastric approach, our preferred method, is not available. The transgastric approach is preferred because the stomach: 1) has a thicker wall, 2) is able to better withstand the forces exerted by the scope, 3) is easier to close, 4) is less likely to form a stricture or obstruction, and 5) allows for repeated access following the placement of an access tube. That said, the creation of an enterotomy access point during this laparoscopic procedure allowed for needle-knife sphyncterotomy to be performed for the treatment of sphincter of Oddi dysfunction.

It is important to stress that adopting the enterostomy route may be challenging in patients with severe adhesions. That was not the case with our patient, in whom all previous operations had been performed laparoscopically. For identification of the afferent limb prior to creation of the enterotomy, we recommend systematically elevating the transverse colon on the left side and identifying the ligament of Trietz. In our experience, this maneuver is much easier than trying to locate the Y-connection and following the bowel backward from that point.

Among the options available for study and therapy of the biliary tree, we considered laparoscopic choledochoduodenostomy and PTC-based approaches. Choledochoduodenostomy would be a valid approach for patients with biliary strictures, but not for our patient who had normal ducts and SOD. In addition, this surgical procedure would introduce the risks of bile leak, biliary duct structuring, and hepatic vasculature injury. A PTC-based approach would have been convenient, since a catheter was already in place. Nevertheless, our suspicion for SOD was very high, and we anticipated performing biliary and pancreatic sphincterotomy, which would not have been possible percutaneously. The pancreatic sphincterotomy was ultimately not performed, as the diagnosis of an internal hernia was thought to explain our patient's abdominal pain, and no history of pancreatitis existed. A biliary sphincterotomy was performed because there was documentation of abnormal LFTs in the setting of pain, supporting the diagnosis of SOD. In retrospect, the PTC examination was not helpful to our patient. Further, the PTC catheter allowed for gas to track into the chest, resulting in a tension pneumothorax. The PTC examination was also the likely culprit of our patient's liver abscesses and right portal-vein thrombosis. PTC catheters frequently cross the diaphragm and render patients susceptible to tension pneumothorax during laparoscopic surgery. Surgeons should be aware of this potential complication and consider the diagnosis when faced with hypotension and hypoxemia in patients with PTC catheters still in place.

Internal herniation is an uncommon, but serious, complication of Roux-en-Y reconstructions. Laparoscopic RYGB affords a shorter hospital stay and decreased postoperative pain, wound complications, and ventral hernia occurrence but carries an increased risk of internal hernia, ranging from 0.2 to 9%.^28,29 Internal hernias can be difficult to diagnose either clinically or radiographically, as patients often present with nonspecific or intermittent symptoms, such as periumbilical pain, nausea, vomiting, anorexia, or abdominal distension. ³⁰ Our patient underwent CT evaluations both at our institution and at the referring hospital, which failed to diagnose the internal hernia. The ability to diagnose and treat internal hernias is an advantage of performing laparoscopic-assisted ERCP over nonsurgical approaches, such as DBE.

Conclusions

Laparoscopy-assisted ERCP procedures are viable in patients with altered anatomy following Roux-en-Y reconstruction. Up to this point, the available literature was limited to the use of the transgastric approach in patients post-RYGB. We present the option of performing laparoscopy-assisted transenterotomy ERCP in patients following partial gastrectomy with Roux-en-Y reconstruction, in whom the transgastric approach is not feasible.