ERCP in surgically altered anatomy requires the endoscopist to fully understand the procedural goals and the reconstructed anatomy before proceeding. Altered anatomy presents a variety of challenges unique to enteroscopy, and others related to accessing the biliary or pancreatic duct from unusual orientations. Both side-viewing and forward-viewing endoscopes, as well as single and double balloon techniques, are available for ERCP in these settings. Endoscope selection largely depends on the anatomy and length of reconstructed intestinal limbs. Endoscopist experience with performing ERCP in surgically altered anatomy is the most important factor for determining outcomes and success rates.
Key points
- •
Performance of endoscopic retrograde cholangiopancreatography (ERCP) in the patient with surgically altered bowel anatomy requires clear understanding of the common surgical rearrangements and thorough knowledge of the patient’s specific surgical history. Endoscope selection is largely based on understanding the patient’s postoperative anatomy, including the presence and lengths of afferent, efferent, or Roux limbs, and the type of biliary drainage present (ie, intact papilla vs bilioenteric/pancreaticoenteric anastomosis).
- •
Alterations that leave the biliary and pancreatic drainage intact at the major papilla (Roux-en-Y gastric bypass, Billroth II procedure, and others) necessitate either traversal of a long limb of gut using an end-viewing instrument, with a difficult approach to the papilla from below, or invasive entry to the excluded stomach to gain traditional access to the papilla from above using a side-viewing duodenoscope.
- •
Alterations that use a surgical biliary or pancreatic anastomosis to jejunum (Whipple procedure; Roux limb hepaticojejunostomy, and so on) generally necessitate long limb access with an end-viewing instrument.
- •
Device selection is largely dependent on the bilioenteric or pancreaticoenteric communication and on the caliber and length of the endoscope selected.
- •
The length and mobility (vs fixation) of bowel needing to be traversed and the endoscopist’s experience are 2 dominant factors related to the overall success of ERCP via altered bowel anatomy.
- •
The success rates, risks, benefits, and planned sequence of alternatives to performance of ERCP in altered anatomy should be thoroughly reviewed with the patient and family. Further studies comparing the different methods for access to the biliary and pancreatic systems are necessary in order to guide clinicians in choosing the most effective, safe, and least costly approach.
Introduction
Since the first reports of endoscopic retrograde cholangiopancreatography (ERCP) in the late 1960s, the procedure has matured into an important yet standard modality for management of pancreatic and biliary disease. Similarly, gastric and biliopancreatic surgery has progressed beyond longstanding procedures for malignant conditions, to commonplace intestinal alterations for liver transplantation and bariatric management. Patients undergoing these various rearrangements are equally or more susceptible to postoperative issues related to biliary stones, dyskinesia, and periampullary neoplasm. Moreover, bilio-enteric and pancreato-enteric anastomoses are subject to stenosis, with associated stone disease or infection. ERCP success rates exceeding 95% are commonplace in patients with normal gastrointestinal (GI) anatomy, whereas patients with surgically altered anatomy pose a greater challenge. In this article, the most common surgical alterations in upper intestinal anatomy ( Table 1 ), and selection of appropriate routes, endoscopes, and accessories for performance of ERCP in these settings, are discussed.
Operation | Biliary and Pancreatic Drainage |
---|---|
| Intact papilla |
Subtotal or total gastrectomy with Roux-en-Y anastomosis | Intact papilla |
RYGB | Intact papilla |
Gastroenteric anastomosis (side-to-side gastrojejunostomy) | Intact papilla |
| Separate new bilioenteric and pancreatoenteric anastomoses |
Roux-en-Y hepaticojejunostomy | New bilioenteric anastomosis |
Cholecystojejunostomy | Pancreatic flow at major/minor papilla intact |
Choledochoduodenostomy | Side-to-side or end to side bilioenteric anastomosis |
Choledochojejunostomy | Biliary continuity at major papilla may be intact as well |
Introduction
Since the first reports of endoscopic retrograde cholangiopancreatography (ERCP) in the late 1960s, the procedure has matured into an important yet standard modality for management of pancreatic and biliary disease. Similarly, gastric and biliopancreatic surgery has progressed beyond longstanding procedures for malignant conditions, to commonplace intestinal alterations for liver transplantation and bariatric management. Patients undergoing these various rearrangements are equally or more susceptible to postoperative issues related to biliary stones, dyskinesia, and periampullary neoplasm. Moreover, bilio-enteric and pancreato-enteric anastomoses are subject to stenosis, with associated stone disease or infection. ERCP success rates exceeding 95% are commonplace in patients with normal gastrointestinal (GI) anatomy, whereas patients with surgically altered anatomy pose a greater challenge. In this article, the most common surgical alterations in upper intestinal anatomy ( Table 1 ), and selection of appropriate routes, endoscopes, and accessories for performance of ERCP in these settings, are discussed.
Operation | Biliary and Pancreatic Drainage |
---|---|
| Intact papilla |
Subtotal or total gastrectomy with Roux-en-Y anastomosis | Intact papilla |
RYGB | Intact papilla |
Gastroenteric anastomosis (side-to-side gastrojejunostomy) | Intact papilla |
| Separate new bilioenteric and pancreatoenteric anastomoses |
Roux-en-Y hepaticojejunostomy | New bilioenteric anastomosis |
Cholecystojejunostomy | Pancreatic flow at major/minor papilla intact |
Choledochoduodenostomy | Side-to-side or end to side bilioenteric anastomosis |
Choledochojejunostomy | Biliary continuity at major papilla may be intact as well |
Terminology
Before performing an ERCP in the patient with surgically altered anatomy, the endoscopist should have a clear understanding of the surgical history of the patient, including the nature and, ideally, the lengths of enteric limbs and the type of pancreatobiliary reconstruction, if any. The terms afferent, efferent, Roux-en-Y, and pancreatobiliary limbs are often used interchangeably and incorrectly. Afferent and efferent describe limb rearrangements in which a gastrojejunostomy, duodenojejunostomy, or other loop-type bypass operation is performed ( Fig. 1 A). The afferent limb refers to the portion of intestine that drains upstream contents (bile, pancreatic juice and other proximal intestinal secretions) toward the gastrojejunostomy; the efferent limb refers to the portion of jejunum that carries these contents downstream and away from the gastrojejunostomy.
The Roux-en-Y operation was named after the Swiss surgeon César Roux, who in 1892 developed a Y-shaped intestinal reconstruction to decompress an obstructed stomach. With this configuration, the jejunum is divided, and the distal limb is brought up to drain an organ or focal pathologic abnormality, such as an obstructed stomach or pancreas, a proximal gastric pouch, the liver, or a pseudocyst. This limb is referred to as the Roux limb (see Fig. 1 B). In most Roux-en-Y applications, the native duodenum and proximal 20 to 40 cm of jejunum are left intact and the jejunal end is anastomosed to the side of the Roux limb 40 to 150 cm downstream from the jejunal disruption, thus forming the opposite limb of the Y from the Roux limb. Whichever limb carries food downstream can be referred to as the alimentary limb (same as the Roux limb in Roux-en-Y gastric bypass [RYGB] anatomy, but the native duodenojejunal limb following Roux-en-Y hepaticojejunostomy). Whichever limb carries the biliary and pancreatic drainage can be called the biliary, pancreatic, or biliopancreatic limb. This limb may be the native upper gut, as in RYGB anatomy or the Roux limb following hepaticojejunostomy or pancreaticojejunostomy. The section of jejunoileum distal to the junction of the 2 Y limbs is termed the common channel. Meticulous use of terminology is important for clinical endoscopy reports, which provide vital information to referring physicians and surgeons.
It is also helpful to distinguish surgical alterations that involve anastomoses of the biliary or pancreatic junction to the intestine (see Table 1 ) from those in which flow through the native papilla is left intact, because this may influence anticipation of success and selection of preferred routes and endoscopic accessories.
Preparation for endoscopic retrograde cholangiopancreatography: a checklist for the endoscopist
Beyond the usual review of indications and appropriateness of ERCP for a given patient, additional considerations ( Box 1 ) include review of prior operative reports, paying attention to the type of reconstruction, length of surgically created limbs of intestine, and types of anastomoses. If certain aspects of the anatomy are unclear, it may be helpful to review the operative report with a surgical colleague or the performing surgeon. All available radiologic studies, such as cross-sectional imaging and upper GI contrast studies, should be reviewed when considering the preprocedure likelihood of a disease process being present and the estimated success rate of reaching, cannulating, and treating the suspected biliary or pancreatic pathologic condition. Most ERCPs performed on patients with surgically altered anatomy are for postoperative biliary issues, such as bilioenteric anastomotic strictures, recurrent cholangitis, and common bile duct stones. Fewer pancreatic interventions have been reported, and interventional procedures are usually in patients who have undergone a pancreatoduodenectomy or Whipple procedure. ERCPs in surgically altered anatomy are often prolonged procedures, so assistance from an anesthesiologist should be considered for the safety and comfort of the patient. Fluoroscopy is often used during these procedures to facilitate deep intubation of long limbs and endoscope tip orientation toward the biliary system. Hence, placement of the patient in a supine or left oblique position can assist with navigation, when appropriate consideration is given to airway protection. Endoscope and accessory selection will be addressed in the discussion of individual anatomic challenges. An overview of endoscope and device compatibilities is available in Table 2 . Working with one’s local representatives for ERCP device manufacturers is another great resource. Finally, the patient and appropriate family members should be thoroughly counseled on not only the goals, risks, and benefits of the procedure but also the overall likelihood of endoscopic success and the various alternatives for management, including observation, referral to another center or endoscopist, other means for access to the papilla, interventional radiology (IR), or surgery.
- 1.
Review operative reports
- •
Understand the postoperative anatomy and lengths of reconstructed limbs
- •
Clarify whether biliopancreaticenteric anastomoses were created
- •
Review postoperative anatomy with surgical colleagues if necessary
- •
- 2.
Review all available radiologic, laboratory, and clinical information
- •
Identify and verify the indications and goals of procedure
- •
- 3.
Select appropriate sedation and patient position
- •
Consider anesthesia-assisted sedation or general anesthesia, given usually prolonged duration of procedure
- •
Supine, prone, or left lateral
- •
- 4.
Select the most appropriate endoscope and accessories
- •
Forward viewing endoscope versus side-viewing duodenoscope
- •
Standard length versus longer endoscope
- •
Single- versus double-balloon overtube versus other accessories to facilitate intubation
- •
Therapeutic (duodenoscope/adult colonoscope) versus standard accessory channel
- •
Straight versus angled catheters and wires, standard versus enteral length
- •
± Clear cap at scope tip
- •
- 5.
Informed consent with patient and family
- •
Goals, risks, benefits of procedure
- •
Honest acknowledgment of endoscopic success rates, alternatives for management of problem (ie, observation, other access to papilla, IR surgery)
- •
Endoscope | Working Length (cm) | Caliber of Biopsy Channel (mm) | Compatible Stents and Accessories | Considerations |
---|---|---|---|---|
Therapeutic duodenoscope | 124 | 4.2 |
|
|
Adult colonoscope | 168 | 3.7 |
|
|
Pediatric colonoscope | 168 | 3.2 |
|
|
Single-balloon enteroscope | 200 | 2.8 |
|
|
Double-balloon enteroscope | 200/220 | 2.8 |
Several key components of ERCP in long limb anatomy directly affect the feasibility and success of the procedure, including the ability to (1) intubate the enteric anastomoses (eg, esophagojejunal, gastrojejunal, or duodenojejunal, or Roux-en-Y entero-enteric); (2) navigate the length of the respective enteral limbs (afferent, Roux, and native duodenojejunal limbs) to reach the papilla or biliary anastomoses; (3) cannulate the respective duct of interest; and (4) accomplish appropriate therapy for the findings identified. Review of a mental checklist or similar process is central to ensuring a safe and successful procedure.
Endoscopic retrograde cholangiopancreatography in altered bowel anatomy without alteration of biliopancreatic anatomy
Gastric Reconstructive Procedures
Partial to complete gastric resection or bypass is often performed for the management of benign and malignant conditions, including the operative treatment of obesity. The 3 most common anatomic rearrangements following gastric resection or bypass are the Billroth I, Billroth II, and Roux-en-Y reconstructions.
Billroth I reconstruction
Theodor Billroth was a leading European surgeon of the nineteenth century who performed one of the first antrectomies in 1881 in which the patient survived. Billroth I and II procedures are performed for reconstruction after partial gastrectomy. With the advent of acid-suppressing therapy and recognition of Helicobacter pylori , they are both far less common than in previous generations. The Billroth I entails a distal gastrectomy/antrectomy, with an anastomosis of the gastric remnant to the duodenum ( Fig. 2 ). This anatomy is the most physiologic type of reconstruction after a distal gastric resection because it restores normal gastroduodenal continuity. Because of the loss of the superior duodenal angle, the duodenum is straightened, and the major and minor duodenal papillae are closer to the stomach. Performance of ERCP in patients with Billroth I anatomy is minimally different from ERCP in those with intact gastroduodenal anatomy. A duodenoscope is still used. Loss of the pylorus and the superior duodenal angle may reduce stability of the endoscope when in the straight or short position, with repeated slippage back toward the stomach. This slippage can usually be overcome by application of slight inward tension on the endoscope, stabilization of intubation depth at the patient’s teeth, or use of a medium to long endoscope position with slight looping in the stomach. Once cannulation has been achieved, all standard therapeutic maneuvers performed in the conventional ERCP without altered anatomy can be performed in the patient with Billroth I reconstruction.
Billroth II reconstruction
The Billroth II reconstruction after a gastric resection is one of the most commonly performed procedures for cancer of the stomach and is also performed for operative treatment of duodenal ulcer disease when a gastric resection is necessary. The distal stomach is resected and the remnant stomach is anastomosed to the proximal jejunum in an end-to-side fashion, with closure of the proximal duodenal stump ( Fig. 3 ). The single gastrojejunal anastomosis at the distal end of the gastric remnant yields entry to side-by-side afferent and efferent lumens with a saddle of jejunal mucosa between them. The orientation of the anastomosis and the position of the jejunal loop in relation to the transverse colon (retrocolic—through the transverse mesocolon vs antecolic—anterior to the transverse colon) are dependent on the patient’s anatomy and pathologic condition and the preference of the surgeon. The Polya method (see Fig. 3 A) uses a wide (>6 cm) anastomosis to the entire gastric opening, whereas the Hofmeister method yields a smaller diameter anastomosis between the greater curvature of the gastric opening and the jejunum (see Fig. 3 B). Sometimes the end of the gastric remnant is closed and a gastrojejunal side-to-side anastomosis is created in the gastric remnant proximal to the closed end of the gastric remnant.
Because of bile reflux into the gastric remnant after a Billroth II reconstruction, a side-to-side jejunojejunostomy between the afferent and efferent limbs is sometimes created to divert the bile away from the remaining stomach. This procedure is known as the Braun procedure ( Fig. 4 ). Endoscopically, this anastomosis will be found in either the afferent or the efferent limb approximately 15 to 20 cm from the gastrojejunostomy. Billroth II anatomy approximates that of the loop gastrojejunostomy performed for distal gastric or proximal duodenal obstruction without resection, so most remarks about afferent limb entry and cannulation in Billroth II anatomy also pertain there.
In Billroth II anatomy, the afferent limb of the duodenum and proximal jejunum often enter the stomach at an acute angle, while the efferent limb of jejunum drains in somewhat more linear fashion from the gastric lumen. Hence, the afferent jejunal limb is more often the leftward oriented lumen on the greater curve or posterior wall, from the patient’s perspective.
Both the forward (gastroscope or colonoscope) and the side-viewing endoscopes (duodenoscope) with a standard or large-diameter accessory channel can be used to achieve and successfully cannulate the biliary tree in the setting of a Billroth II reconstruction. Overtubes or smaller channel enteroscopes are rarely needed. Papillary cannulation is facilitated by use of the duodenoscope, which is the preferred instrument in experienced hands, but in many circumstances the length or fixation of the afferent limb requires use of more flexible or longer end-viewing instruments. Larger series studies of ERCP in Billroth II anatomy have demonstrated success rates using duodenoscopes for reaching the papilla and for selective biliary cannulation of 70% to 97% and 60% to 91%, respectively. Entry from stomach to the afferent limb carries a risk of anastomotic or jejunal perforation, particularly with torque entry of the duodenoscope. Advantages of forward-viewing endoscopes, most of which are less rigid than duodenoscopes, include greater ease of entry to the afferent loop and safety of deep intubation to the level of the papilla. Cannulation rates as high as 81% to 87% have been reported using forward-viewing endoscopes in this patient population. Forward and side-viewing endoscopes were compared in a prospective, randomized trial of ERCP in 45 patients with Billroth II anatomy. Cannulation rates were higher (87% vs 68%) with the use of gastroscopes. However, these results must be interpreted with caution, as all cannulation failures in the duodenoscope group (N = 7) were due to the inability to reach the papilla as a result of intestinal perforations, difficulty in entering the afferent limb, or abdominal pain during insertion. Therefore, factors related to endoscopic skills and comfort level of the endoscopist as well as anatomic factors related to the surgical reconstruction will guide the endoscopist in selecting the appropriate endoscope.
The afferent and efferent limbs can often be differentiated by visualization of bile and observation of motility on initial entry. The afferent limb should contain more bile, with motility propulsing toward the lens. The proximal end of the afferent limb is characterized by a blind stump of flat mucosa, with or without an obvious closure line, scars, staples, or sutures. Fluoroscopy can help clarify the direction of the endoscope if the correct lumen or extent of intubation is in question. Demonstration of the endoscope in, or heading toward, the right upper quadrant and the liver suggests entry into the afferent limb. Evidence of a biliary aerogram is uncommon with an intact papilla but may prove useful. Advancement of a slippery wire and a biliary occlusion balloon ahead of the endoscope can confirm the correct limb while also characterizing and facilitating advancement through acutely angulated loops of jejunum. Occasionally repositioning the patient in the supine position allows hand compression over the abdomen to reduce loops formed by the endoscope.
Once in the duodenum, the duodenoscope provides a superior view of the papilla from below, compared with the eccentric view seen with forward-viewing endoscopes. Availability of an elevator also facilitates device manipulation, although overuse contributes to inadvertent pancreatic cannulation. The major papilla is commonly found near the 12 o’clock position of the visual field when a duodenoscope is used in a patient in the left lateral decubitus position, and eccentrically at the 3 to 6 o’clock position when a forward-viewing endoscope is used. Visualization and approach to the papilla from below with any end-viewing instrument are facilitated by use of a clear plastic cap affixed to the endoscope ( Fig. 5 , [CR] ). Compression of the duodenal angle between the second and third portion and of periampullary folds enables a slightly more distant and useful view of the papilla than can be achieved with a bare endoscope tip.
When viewed from below using a duodenoscope, the paths of the bile duct and the pancreatic duct can be confusing, because they are upside down from standard views seen with normal antegrade positioning of the side-viewing instrument. The apex of the ampulla and the upward directed bile duct lie at 5 to 6 o’clock in the visual field, rather than the usual 12 o’clock. The pancreatic duct penetrates from the midpoint of the papilla at a slight angle. Hence, standard cannulas with an upward curve toward 12 o’clock in the visual field will orient predominantly into the pancreas. This dilemma can best be overcome by retracting the lens to the angle between second and third portion and viewing the papilla from a distance while passing a straightened catheter and a nonangled guidewire directly upward in a vertical axis aligned with the bile duct, avoiding or limiting use of the elevator, which redirects devices into the pancreas.
Once into the bile duct, cholangiography is straightforward, but therapy, beginning with performance of sphincterotomy, requires accommodation to the upside down view. A variety of inverted, upside down, and “shark-fin” sphincterotomes have been described, but sphincterotomy is generally most safely and easily accomplished by insertion of a 5- or 7-Fr temporary biliary stent for use as a guide to needle-knife incision. Success rates as high as 95% have been reported with the over-the-stent needle-knife technique in one series, and with a low complication rate (5%). Free-hand needle-knife sphincterotomy is occasionally required if cannulation cannot be achieved; however, this approach carries increased risk of periampullary perforation. Balloon sphincteroplasty, or dilation of the intact or partially incised sphincter, can reduce the risk of full sphincter incision, but dilation of an intact papilla heightens the risk of pancreatitis in some patients. Balloon dilation catheters do not require changing the angle used to approach the papilla, and furthermore, usually come in a sufficient length for use with even longer endoscopes, such as single- and double-balloon enteroscopes, making them widely available in most endoscopy suites. In addition, balloon dilation may be more appealing rather than sphincterotomy in patients with increased risk of bleeding. In a single-center study in which 34 patients with bile duct stone and a previous Billroth II gastrectomy were randomized to ERCP with a side-viewing duodenoscope and either endoscopic sphincterotomy (EST) or endoscopic balloon dilation (EBD), there were no statistically significant differences in the rates of bile duct stone removal (14/18 patients in EST group vs 14/16 in EBD group, P = 1.00), use of mechanical lithotripsy (4 EST procedures vs 3 EBD procedures, P = 1.00), or early complications (7 EST procedures vs 3 EBD procedures, P = .27). Three patients in the EST group experienced bleeding requiring blood transfusions, and one patient in the EBD group developed mild pancreatitis. Thus, although both EST and EBD are options, both carry a small, yet present, risk of complications that are unique to each technique. Additional studies have shown satisfactory cannulation rates with balloon dilation alone, with either duodenoscope or forward-viewing endoscopes, with success rates ranging from 83% to 92%. Dilation of the papilla combined with guidewire-assisted needle-knife papillotomy is another option that has also been described as successful.
Subtotal gastrectomy or gastric bypass with Roux-en-Y anastomosis
All Roux-en-Y anastomoses to the esophagus or proximal gastric remnants, whether performed with extended gastric resection or gastric bypass, pose similar challenges for performance of ERCP. Total or subtotal gastrectomy with Roux-en-Y anastomosis is usually performed as a primary oncologic procedure, or in some circumstances, as salvage of a complicated gastric bypass. Roux limb anastomosis to more generous proximal gastric segments is occasionally done in revision of a previous Billroth resection to reduce troublesome bile reflux gastritis or esophagitis. The length of a nonbariatric alimentary Roux limb is typically at least 40 cm in length from the esophagojejunal or gastrojejunal anastomosis to the jejunojejunal anastomosis at the base of the Y ( Fig. 6 ). Hence, no bile should be present in the remnant stomach or esophagus, regardless of patency of the biliary tree.