ERCP and Related Technologies: A Clinical Update

David L. Carr-Locke, MD, FRCP, FACG, Brigham and Women's Hospital and Harvard Medical School, Boston, Massachusetts.


May 22, 2000

In This Article

Sphincter Disease

This session covered complications of endoscopic sphincterotomy, palliation of ampullary tumors, and sphincter of Oddi disease.

Dr. Chung, from the Chinese University of Hong Kong, discussed endoscopic sphincterotomy as a technically demanding procedure performed in the most hazardous part of the gastrointestinal tract. "Complications occur in around 10% of patients with a procedure-related mortality of about 0.5%[1] and may be classified as mild (2-3 days in hospital), moderate (4-10 days in hospital), or severe (more than 10 days in hospital, need for surgical or radiologic intervention, or contributing to death).[2]"

He reported that oxygen desaturation to dangerously low levels occurs frequently during ERCP[3] and recommended preventing this by giving oxygen by naso-cannula at a rate of 4.5 L/min4.

Pancreatitis is the most common complication after sphincterotomy, occurring in around 5% of cases, and is more likely in sphincter of Oddi dysfunction, younger patients, difficult cannulation, and precut techniques.[1] Once pancreatitis has occurred, no specific measure has been shown to improve the outcome.

Bleeding occurs in about 2% of patients, with increased risk in coagulopathy, in patients taking anticoagulants, cholangitis, when there is bleeding during the procedure,[1] or after "zipper" sphincterotomy. This complication can be avoided by using a "smart" electrocautery generator, such as the ERBE Endocut[5]; or by avoiding sphincterotomy in high-risk cases altogether and relying on balloon dilatation of the sphincter.[6] Treatment of postsphincterotomy bleeding is by flushing with 1/100,000 epinephrine, followed by injection of 1/10,000 epinephrine if bleeding continues. In a series of 982 sphincterotomies, bleeding was reported in 12%, with 46% responding to epinephrine flushing and the remainder to injection. Rebleeding occurred in only 5 patients, all responding to repeat injection.[7] Multipolar coagulation has also been used successfully.[8] Angiographic embolization of the appropriate branch of the gastroduodenal artery is the next step if bleeding continues,[9] and surgical sphincteroplasty is a last resort.

Perforations are rare, occurring in only 0.3% of cases.[1,10] Guidewire perforations are usually benign, but Dr. Chung recommended antibiotics for a few days. Intraperitoneal perforations caused by the endoscope are uncommon, but are more likely in patients with surgically altered anatomy, such as the afferent loop of the B II partial gastrectomy.[11] Surgical repair is indicated. Retroperitoneal perforation from the sphincterotomy site should be treated conservatively initially, but management is controversial.[12,13,14] If continuing pathology requires treatment, such as a bile duct stone, early surgery is indicated; otherwise, a nasobiliary drain is advised, nasogastric suction, and antibiotics, with surgery indicated if abdominal signs do not respond rapidly in 24 hours. As a surgeon himself, Dr. Chung advised surgeons not to search for the perforation, but to drain the bile duct and the retroperitoneal space.

Post-ERCP sepsis could be precipitated by forcible injection of contrast, which could be avoided by religiously aspirating bile before injecting. Adequate biliary drainage is essential to avoid delayed cholangitis in an obstructed system and stents or nasobiliary tubes are equally effective.

Dr. Deviere, from Erasme Hospital in Brussels, quoted 3- and 5-year survival rates for ampullary carcinoma of 50% and 40%, respectively,[15] with the mortality from pancreaticoduodenectomy as low as 2% in experienced centers, mitigating in favor of surgery in every fit patient. Palliation was restricted to patients in poor condition and those with nonresectable tumors. Endoscopic techniques included tumor sphincterotomy, ampullectomy, and stenting.

With respect to villous adenomas of the papilla,[16] resection is the primary goal, and, if feasible, endoscopic snare resection with careful follow-up seems to be justifiable, as in the series from Hamburg reported in 1993.[17] Such an approach is probably preferred to thermal ablation or PDT,[18] because it provided pathology and resection margin.

Dr. Lehman defined sphincter of Oddi dysfunction (SOD) as a benign, noncalculus obstructive disorder of the sphincter of Oddi causing pancreatobiliary pain, cholestasis, and/or pancreatitis. The incidence of SOD is uncertain, because different groups are constrained by selection bias and referral patterns. Rates vary from 9%[19] to 11%[20] to as high as 62%[21] when based on abnormal sphincter of Oddi (SO) manometry, which is present in 72%, 62%, and 60% of types I, II, and III disease if biliary and pancreatic sphincter pressures are recorded.[21] Pathologic changes in the biliary sphincter segment have been shown in surgical specimens[22,23] in 60% of cases, with the remainder representing a motor disorder. Although neural control of the sphincter segment is thought to be from nonadrenergic, noncholinergic nerves in which vasoactive intestinal peptide (VIP) or nitric oxide (NO) is the neurotransmitter, there are as yet no useful observations in this area to assist with diagnosis or therapy. Dr. Lehman summarized his view on noninvasive diagnostic tests for SOD as being generally unhelpful in the less obvious cases, but useful when invasive tests are to be avoided.[24,25,26,27,28,29]

Although invasive and difficult to perform, SO manometry has become the standard for diagnosis of SOD,[30,31] with an upper normal value for basal sphincter pressure of 40 mm Hg based on results in asymptomatic controls.[32] For SO manometry, all drugs having any effect on the sphincter should be avoided for up to 12 hours prior to the examination, although current evidence suggests diazepam, droperidol, and meperidine are acceptable for sedation. A 5-French triple lumen catheter has become standard, and some can be passed over a guidewire and the duct identified by aspiration. Dr. Lehman stated that, ideally, both pancreatic and biliary pressures should be studied, because abnormalities may be confined to 1 side of the sphincter in 35% to 65% of patients.[21,33,34] An aspiration catheter reduced postmanometry pancreatitis dramatically for his group,[35,36] and he recommended its routine use.

Therapy of SOD remains controversial, except in cases of type I obstruction, for which endoscopic sphincterotomy or surgical sphincteroplasty are both effective. These techniques in lesser degrees of obstruction require careful risk assessment in view of the known increased complication rate from endoscopic sphincterotomy (which is up to 4 times more than that among patients with stone disease).[37] Randomized controlled trials have shown a convincing response to endoscopic sphincterotomy compared with sham therapy in type II patients, with abnormal manometry in 50% to 90%.[21,22,38,39] The patients with type III obstruction remain an enigma. Recent experience with Botox injection into the sphincter as a predictor of a good outcome from sphincterotomy[40] is encouraging, but indiscriminate sphincterotomy in these patients should be balanced by the poor response rate (only 50%) and high complication risk.[37,41] Pancreatitis can also be reduced by a protective pancreatic stenting prior to sphincterotomy, as demonstrated in 3 of 4 prospective randomized trials.[42,43,44,45] Although medical therapy has been disappointing, with a limited literature to support it, Dr. Lehman advised that it be tried first in all patients before resorting to more invasive therapy.[46,47]

Lastly, Dr. Lehman addressed the issue of nonresponders after biliary sphincterotomy and classified them into 4 groups: incomplete biliary sphincterotomy, residual pancreatic sphincter dysfunction,[48] unsuspected presence of chronic pancreatitis,[49,50] and an alternative diagnosis, such as generalized gut dysmotility.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: