Key points

Introduction

Endoscopic retrograde cholangiopancreatography (ERCP) was first performed in 1968 and the first sphincterotomy was carried out in 1974. ERCP has evolved into a routine endoscopic procedure, with approximately 500,000 procedures performed in the United States and 1.3 million worldwide annually. Most procedures in the United States are therapeutic. ERCP represents a monumental advance in the management of patients with pancreaticobiliary diseases, but is a complex and technically demanding procedure with the highest inherent risk for adverse events of all routine endoscopic procedures. This article evaluates patient-related and procedure-related risk factors for ERCP-related adverse events, and discusses strategies for the diagnosis and management of these events.

Introduction

Endoscopic retrograde cholangiopancreatography (ERCP) was first performed in 1968 and the first sphincterotomy was carried out in 1974. ERCP has evolved into a routine endoscopic procedure, with approximately 500,000 procedures performed in the United States and 1.3 million worldwide annually. Most procedures in the United States are therapeutic. ERCP represents a monumental advance in the management of patients with pancreaticobiliary diseases, but is a complex and technically demanding procedure with the highest inherent risk for adverse events of all routine endoscopic procedures. This article evaluates patient-related and procedure-related risk factors for ERCP-related adverse events, and discusses strategies for the diagnosis and management of these events.

General considerations

Before considering performing an ERCP, the endoscopist must ensure there is an appropriate indication. There has been a dramatic reduction in diagnostic ERCPs given the increased use of less invasive imaging modalities to view the pancreaticobiliary system (eg, MRI, computed tomography, and endoscopic ultrasonography [EUS]). Endoscopists must also consider if they have adequate volume to maintain their ERCP skills and minimize adverse events. Overall adverse event rates for ERCP are typically reported as 5% to 10%. The most commonly reported complications include post-ERCP pancreatitis (PEP), bleeding, perforation, infection (cholangitis), and cardiopulmonary or sedation-related events. This article reviews these adverse events and its companion article elsewhere in this issue by Drs Rustagi and Jamidar focuses solely on PEP.

Post–endoscopic retrograde cholangiopancreatography pancreatitis

Given the importance of this topic and the large amount of published data, PEP is covered in greater detail in the accompanying article by Drs Rustagi and Jamidar focuses on PEP. This section provides a brief overview of the subject.

PEP is one of the most common and feared adverse events of ERCP, resulting in considerable morbidity and, rarely, death. Several patient-related (younger age, female gender, history of previous PEP, nondilated ducts, normal bilirubin level, suspected sphincter of Oddi dysfunction) and procedure-related factors (difficult cannulation, multiple pancreatic injections, pancreatic sphincterotomy, precut sphincterotomy, pancreatic sampling) have been identified as increasing the risk of PEP. Therefore, careful patient selection and sound endoscopic technique are the cornerstones in the prevention of PEP.

Prophylactic pancreatic duct stenting, particularly in high-risk patients, has been shown to reduce the incidence and severity of PEP by mechanically facilitating pancreatic drainage.

In addition, chemoprophylaxis of PEP has been extensively researched in an attempt to prevent or reduce the severity of PEP. Numerous trials studying a variety of pharmacologic agents (eg, somatostatin, octreotide, corticosteroids, allopurinol, protease inhibitors, nitroglycerin) have yielded disappointing results ; however, an important exception is rectal administration of nonsteroidal anti-inflammatory drugs (NSAIDs), which have been shown in large randomized controlled trials and meta-analyses to significantly reduce the incidence and severity of PEP.

Bleeding

Bleeding observed at the time of or after ERCP is usually related to endoscopic sphincterotomy. Bleeding seen endoscopically during or immediately after sphincterotomy is not uncommon, but is generally not considered an adverse event unless there is clinically significant blood loss, transfusion, or a major change in management.

The true incidence of significant ERCP-related hemorrhage is variable and difficult to define given the lack of a standard definition. Some degree of bleeding (ranging from oozing to severe bleeding) is seen at the time of sphincterotomy in up to 10% to 30% of cases. However, endoscopists are generally unconcerned about minor bleeding, such as limited oozing immediately after a sphincterotomy, because in most cases this is temporary and stops spontaneously. When applying clinical criteria such as melena, hematemesis, a greater than 2 g/dL drop in hemoglobin level, or requirement for secondary intervention such as endoscopy or blood transfusion, the overall incidence of bleeding is around 0.1% to 2%. Based on severity, bleeding has been classified into 4 groups: (1) mild (clinical evidence for bleeding but drop in hemoglobin is <3 g/dL; no blood transfusions), (2) moderate (endoscopic treatment required; transfusion requirement ≤4 units), (3) severe (transfusion of ≥5 units and/or surgery or angiographic treatment), and (4) fatal.

Immediate bleeding, defined as continued bleeding 2 to 3 minutes after the initial sphincterotomy, is seen in 50% to 60% of patients. Delayed bleeding, defined as occurring after the completion of ERCP, can happen hours or up to 7 to 10 days after the procedure.

Several risks for postsphincterotomy hemorrhage have been identified, which allow risk-stratification and risk-reduction measures to be taken before the procedure. Definite risk factors include any degree of bleeding during the procedure, presence of any coagulopathy or thrombocytopenia, initiation of anticoagulant therapy within 3 days after the sphincterotomy, presence of active cholangitis, and endoscopist’s low case volume (performance of <1 sphincterotomy per week). Other potential risk factors include ampullary stone impaction, periampullary diverticula, uncontrolled cutting (the so-called zipper cut), and needle-knife sphincterotomy. Factors that do not seem to increase risk of bleeding include longer sphincterotomy incision, extension of a previous sphincterotomy, and use of aspirin or NSAIDs. The risk of bleeding with newer agents including clopidogrel is unclear.

In patients with thrombocytopenia or coagulopathy, risk can be minimized by transfusion of blood products (with goal of platelet counts >50,000/mm ³ and international normalized ratio <1.5–2), withholding anticoagulants for 3 days afterward, and avoiding sphincterotomy by use of alternative procedures such as balloon sphincteroplasty. Prophylactic epinephrine injection into the sphincterotomy site in high-risk patients has been suggested to reduce the risk of postsphincterotomy bleeding but is of uncertain efficacy. Intraprocedurally, the use of blended current through automated current delivery cautery systems has been shown to reduce the risk of immediate bleeding.

Many types of endoscopic therapies are available for immediate or delayed postsphincterotomy hemorrhage. Injection of dilute epinephrine (1:10,000) is the most common first-line endoscopic approach. Varying amounts are injected at a targeted site or in the apex of the sphincterotomy site if bleeding obscures visualization. Needles with a metal shaft and spring sheath (Carr-Locke injection needle; US Endoscopy, Mentor, OH, USA) may pass more easily through the tip and elevator of the side-viewing duodenoscope. A comparison between epinephrine injection monotherapy and combination endotherapy (epinephrine plus cautery) was retrospectively evaluated in a single-center experience, resulting in similar rates of success (96.2% vs 100%; P = .44), recurrent hemorrhage (16% vs 12.1%; P = .72), other procedural adverse events, transfusions, angiographic embolization, and surgery.

Balloon tamponade using a standard stone extraction or dilation balloon may allow control of bleeding and improve visualization of the bleeding site. Thermal coaptive coagulation using either a multipolar probe or heater probe device can follow, particularly if a specific bleeding point can be identified. Endoscopic clip placement is another option, and can be very effective. Caution should be taken to avoid thermal injury or clip placement over the pancreatic duct orifice, especially if the bleeding site is to the right of the sphincterotomy incision contiguous to the pancreatic duct orifice. Fully covered self-expandable metallic stents (FCSEMS) have also been used with success to create a tamponade effect at the area of bleeding as a rescue technique if other endoscopic methods have failed. Application of a catheter-delivered hemostatic powder, TC-325 (Hemospray; Cook Medical, Winston-Salem, NC, USA), is a promising potential therapy for hemostasis of sphincterotomy-induced hemorrhage but is not currently available in the United States. Rarely, angiographic embolization or surgery is required for refractory bleeding.

Perforation

Perforation is an uncommon adverse event, occurring in 0.3% to 0.6% of cases. Three distinct types of perforations can occur: (1) free wall perforation of the esophagus, stomach, or duodenum by the endoscope, resulting in mediastinal or intraperitoneal leakage; (2) retroperitoneal perforation as a result of extension of a sphincterotomy incision beyond the intramural portion of the bile or pancreatic duct; and (3) perforation of the bile or pancreatic duct from extramural passage of guide wires or stents.

Bowel wall perforations are force and angle related, and are more likely to occur in patients with esophageal stricture, Zenker diverticulum, postsurgical altered anatomy with fixed angulation, or gastric outlet obstruction caused by advanced pancreatic cancer. The duodenum is the most common site of free bowel wall perforation, which typically occurs with mechanical pressure from a rigid duodenoscope tip tearing a thin-walled proximal duodenum, or when the scope tip tears a periampullary diverticulum. Sphincterotomy-related perforations are more common after needle-knife precut access, and in patients with suspected sphincter of Oddi dysfunction. Guide wire–related ductal perforations typically occur through a side branch of the pancreatic duct or hepatic capsule, but have become less common with increasing use of floppy-tipped, hydrophilic wires. Dilation of biliary or pancreatic duct strictures and the use of a large extraction balloon in a small-caliber duct can also lead to intraductal rupture.

Treatment of post-ERCP perforation depends on the type, size, and location of the perforation and severity of the leak, clinical manifestations and stability of the patient, and available devices and expertise. This topic is also covered in the article discussing perforation closure and management elsewhere in this issue. Early recognition and endoscopic treatment of suspected perforation along with conservative management has been shown to have favorable outcomes. Although free bowel wall perforations have been managed historically with surgical repair, improved endoscopic closure devices have permitted nonsurgical management in an increasing number of cases, particularly if the perforation is recognized at the time of the endoscopy.

Attempts to close perforations endoscopically should be performed using CO ₂ to minimize the risk of tension pneumoperitoneum or extensive retroperitoneal, mediastinal, and subcutaneous gas accumulation, including pneumothorax. Several reports supporting the primary endoscopic closure techniques using an endoscopic clip, endoloop, or the over-the-scope clip (OTSC) to treat larger perforations have been described, even for use in direct perforation of the duodenal wall. The OTSC (Ovesco Endoscopy AG, Tübingen, Germany; Padlock, Aponos Medical, Kingston NH, USA), have been successfully used for closure of esophageal, gastric, and duodenal perforations. The OTSC enables more durable closure than the through-the-scope clip because of its ability to grasp more tissue (by pulling the defect edges into the cap before clip deployment) and apply a greater compression force for closure. Endoscopic suturing (OverStitch; Apollo Endosurgery, Austin, TX, USA) is an option for accurate endoscopic full-thickness closure of iatrogenic perforations, though suturing is limited by the availability of the device, length of the dual-channel upper endoscope, and difficulty in operating within the narrow duodenal lumen. In addition, FCSEMS have also been used successfully to close esophageal and, more specifically, scope-induced duodenal wall perforations.

Wire-related ductal perforations can usually be treated endoscopically by inserting a stent in the respective duct and providing transduodenal duct drainage beyond the leak site. Sphincterotomy-related perforation may be minimized by limiting the length of cutting wire in contact with the tissue, using stepwise incisions, and using the newer-generation microprocessor-controlled electrosurgical generators. The safety profiles of cutting currents provided by these generators have eliminated any need to use a pure cut setting. If perforation is suspected during a sphincterotomy, careful fluoroscopy and injection of a small amount of contrast while pulling the papillotome through the incision over a guide wire can confirm or exclude extravasation. Endoscopic clipping may be attempted to close a definite leak. An FCSEMS can also be placed to drain the bile duct and occlude the leak site in biliary sphincterotomy-related perforation.

In addition to endotherapy, patients should be admitted, kept nil by mouth with nasogastric suction, given intravenous acid-suppressive therapy and intravenous antibiotics, and have a surgical consultation. Furthermore, abdominal computed tomography should be obtained to assess for the degree of contrast leakage, fluid collection, and inflammation. If there is significant leak with ongoing contrast extravasation, or deterioration in a patient’s clinical condition, prompt surgical or percutaneous drainage is recommended.