GI Consult: Acute Pancreatitis

An update on predicting the course of the disease, recognizing clues to pancreatitis of biliary orgin, and deciding when endoscopic examination is warranted.

By John Baillie, MB, ChB

1. What is acute pancreatitis?

Acute pancreatitis is a clinical syndrome consisting of epigastric abdominal pain, often radiating to the back; nausea; vomiting (not always present); and a serum amylase or lipase level greater than three to five times normal. The condition is frequently misdiagnosed. Hyperamylasemia, for example, is not synonymous with acute pancreatitis. An elevated serum amylase level can be triggered by numerous nonpancreatic causes, including perforated duodenal ulcer, ruptured ectopic pregnancy, and small bowel obstruction. Likewise, an elevated pancreatic enzyme level can be caused by noninflammatory disorders, such as pancreatic tumor or disruption of the pancreatic duct caused by sharp or blunt trauma. The diagnosis of acute pancreatitis depends not only on the clinical presentation but also on serologic evidence of abnormal pancreatic function or confirmatory results obtained from a cross-sectional pancreatic imaging study, such as a contrast-enhanced computed tomography (CT) scan.

2. Can acute pancreatitis be accompanied by normal serum amylase and lipase levels?

Yes, but such a finding would be unusual in all but the mildest cases. Amylase is a small molecule that is rapidly and completely cleared by the kidneys. The serum amylase level may be normal if it is measured more than 24 to 48 hours after onset of symptoms. Urinary amylase used to be measured as a way to detect an increase in serum amylase concentration, since urinary excretion trails serum levels by 6 to 12 hours or more. The serum lipase level, however, rises more slowly and persists for up to a week after the onset of acute pancreatitis, making it a better historic marker for this disorder. Many clinicians have stopped relying on serum amylase levels and instead favor serum lipase estimations for detecting acute pancreatitis.

Other serologic indicators of acute pancreatitis include C-reactive protein (CRP), leukocyte elastase, trypsinogen-activating peptides, and lactate dehydrogenase (LDH). Few clinicians routinely perform tests for these markers, however. The urinary trypsinogen-II assay is a rapid (three-minute) dipstick test performed at the bedside that is highly sensitive and specific in detecting acute pancreatitis. Unfortunately, this test is not currently licensed for use in the U.S. Serologic tests that can detect the proinflammatory cytokines, such as interleukins, that are associated with acute pancreatitis are in development.

3. Should every patient presenting with acute pancreatitis undergo abdominal CT scanning?

No. Computed tomography scanning should be used only when the diagnosis is uncertain, the pancreatitis is expected to be severe, or the disorder worsens or fails to resolve. Normal results from an abdominal CT scan suggests that acute pancreatitis is not likely. Even mild disease will produce telltale signs, such as stranding of the peripancreatic tissues, caused by inflammatory edema, and sometimes fluid collection. Because a good-quality CT scan will illustrate many of the nonpancreatic causes of abdominal pain and an elevated serum amylase level, it can serve as an excellent screening test in cases of diagnostic difficulty.

Most patients with predicted severe acute pancreatitis will have pancreatic necrosis. This disorder will be evident during an intravenous contrast-enhanced CT scan, because the pancreatic parenchyma will not turn opaque during the arterial phase of contrast injection. When less than 50% of the pancreas turns opaque during the procedure, a prolonged and often turbulent course is likely and will almost inevitably include percutaneous or surgical drainage or removal of collected fluid and necrotic tissue.

Some patients who have less severe pancreatitis may not show signs of improvement, or their condition may worsen, as indicated by increasing pain, fever, and a high white blood cell count. In these patients, an unsuspected sequela of the disease may be to blame, such as fluid collection, which is often the product of inflammatory disruption of the pancreatic duct. Abdominal CT scanning is therefore recommended in such cases for detecting these complications.

4. What role does transabdominal ultrasonography (TUS) play in detecting the complications and sequelae of acute pancreatitis?

Transabdominal ultrasonography is a sensitive test for detecting cholelithiasis, as well as for detecting dilatation of the extrahepatic biliary tree, which may indicate the presence of choledocholithiasis. In detecting bile duct stones, TUS is more than 95% specific but only about 60% sensitive. In patients with acute pancreatitis the pancreas itself is rarely well visualized by TUS, because air is usually present in the distended loops of the small bowel. Transabdominal ultrasonography is inexpensive, portable, and readily available in most hospitals. However, the quality of the results depends significantly on the thoroughness of the scan and on the training and experience of the technician. When TUS is performed on patients who may have acute pancreatitis, the operator should be asked to scan the entire abdomen, as a nonbiliary, nonpancreatic disorder, such as an ovarian cyst or blood in the pelvis, may be the cause of abdominal pain.

5. How does magnetic resonance cholangiopancreatography (MRCP) compare with abdominal CT scanning?

Magnetic resonance imaging (MRI) of the biliary tree and pancreas has developed rapidly over the last five years, but its role in detecting the complications and sequelae of acute pancreatitis has yet to be defined. The technique could serve as an alternative to abdominal CT scanning for patients who have renal impairment or a history of severe reaction to intravenous contrast media. Unfortunately, the sickest patients with acute pancreatitis are the least likely to undergo MRI, because they are usually treated in an intensive care unit (ICU), attached to mechanical ventilators and other medical machines. In addition, the sensitivity of MRI in detecting pancreatic necrosis has yet to be determined.

These drawbacks limit the utility of MRCP during the acute phase of pancreatitis. However, the test can help locate the causes of unexplained, recurrent pancreatitis, including pancreatic ductal disruption, fluid collections, and occult tumors.

6. Can the course of acute pancreatitis be predicted with any accuracy?

The answer to this is a guarded yes. A variety of scoring systems have been developed over the years to determine which patients are likely to suffer the sequelae of acute pancreatitis. Ranson's criteria, for example, assess 11 risk factors, including a patient's age, blood glucose, and serum calcium and blood urea nitrogen (BUN) levels, which are determined at the time of admission and at 24 to 48 hours afterward. Each criterion the patient fulfills adds one point to the score, which is related linearly to the risk of death. A score of 8 to 11 indicates a high risk (greater than 50%), usually attributed to systemic complications such as circulatory collapse, renal failure, and acute respiratory distress syndrome (ARDS).

The Ranson's criteria requirement for a second assessment at 24 to 48 hours after the first is an indication of how long the systemic complications of acute pancreatitis take to develop. Such a period is considered to reflect the delayed effects of circulating proinflammatory cytokines, such as tumor necrosis factor-alpha (TNF-a), interleukins, and platelet aggregating factor (PAF).

Since the Ranson's criteria were developed, clinicians have devised other, less complex scoring systems, such as the Glasgow (Imrie) criteria, with the purpose of simplifying the initial assessment. However, no special assessment is necessary for a physician to realize that a patient who is hypotensive, hypoxic, oliguric, febrile, and confused is seriously ill. The APACHE scoring systems II and III currently in routine use in ICUs provide similar results in predicting outcome, but they are cumbersome and take longer to provide the desired prediction.

Patients with predicted severe acute pancreatitis are most appropriately cared for in an ICU or other setting where their rapidly changing status can be continuously monitored and medical staff can respond immediately to any metabolic derangements. Does every patient with a high Ranson score develop severe pancreatitis? No. Those of us who treat this disorder have all seen similar patients who have surprised us with an uneventful course and recovery. However, these patients are the exception.

Finally, as previously mentioned, contrast-enhanced abdominal CT scanning is both sensitive and specific for acute pancreatitis and pancreatic necrosis. The Balthazar Score is a predictive system based on CT findings that is probably as accurate as any of the other scoring systems.

7. Should patients who have gallstone pancreatitis undergo urgent endoscopic retrograde cholangiopancreatography (ERCP)?

This is a very important and controversial question. Removing a common bile duct stone endoscopically to reverse or abort the clinical course of acute, gallstone pancreatitis is an attractive option. Frequently, endoscopists are asked to perform urgent ERCP to extract endoscopic bile duct stones from patients who have acute pancreatitis of suspected biliary origin.

In the absence of other obvious causes of acute pancreatitis, especially when a patient is known to have gallstones, the diagnosis of biliary pancreatitis is usually straightforward. Elevated serum liver measurements, especially an alanine aminotransferase (ALT) level greater than three times normal, that are not explained by another obvious cause are suggestive of biliary pancreatitis, as are an elevated liver enzyme concentration, jaundice, and a dilated extrahepatic bile duct, with or without stones. In particular, if a patient has predicted severe acute pancreatitis, the endoscopist may be asked to perform urgent ERCP to prevent worsening of the disease.

So far, three prospective, randomized trials have compared early endoscopic intervention (ERCP/sphincterotomy) with conservative methods in the treatment of gallstone pancreatitis. In their separate studies of patients who had acute pancreatitis accompanied by biliary obstruction, Neoptolemos, Fan, and their colleagues observed a reduced incidence of cholangitis and a trend toward early ERCP.

In the third study, Folsch and colleagues excluded patients who had biliary obstruction and observed the effect of urgent ERCP on patients who had acute pancreatitis alone. Their results suggest that the procedure not only provides no benefit to patients who show no evidence of biliary obstruction (jaundice with or without cholangitis) but also may produce complications that worsen the disease.

In a large study of Polish patients, Nowak and colleagues report evidence strongly in favor of performing early ERCP for all patients who have gallstone pancreatitis. However, because the researchers' methods and interpretation of the data have since been questioned, their findings must be considered inconclusive.

The soon-to-be-published results of an important study conducted by the Virginia Mason Clinic in Seattle, Washington, and presented at Digestive Diseases Week (DDW) in May 2000, suggest that ERCP may have a nonbiliary role in the evaluation of patients who have predicted severe acute pancreatitis. Of 100 patients who underwent early ERCP, 75 had disruption of the main pancreatic duct and subsequently underwent several different types of treatment ranging from endoscopic stent implantation to surgery. Until the data from this study are published, the consensus among experts is that for now, only patients who have progressive biliary obstruction, as evidenced by progressive jaundice with or without cholangitis, should undergo urgent ERCP for biliary compression.

The stones that cause gallstone pancreatitis are small, usually 5 mm or smaller in diameter, and most of those pass spontaneously. Within 48 to 72 hours after the onset of gallstone pancreatitis, the likelihood that ERCP will detect a bile duct stone is less than 20%.

8. Patients who recover from gallstone pancreatitis frequently undergo laparoscopic cholecystectomy before leaving the hospital. Should these patients all undergo preoperative ERCP to "clear" the bile duct?

No. This step is unnecessary and costly and increases the risk of complications associated with ERCP. As noted above, most patients who have had gallstone pancreatitis pass their stones spontaneously. It is therefore common in these patients for serum liver concentrations of bilirubin and transaminase and other enzymes to be mildly elevated, which slowly normalize. These patients can and should undergo cholecystectomy—usually laparoscopic cholecystectomy—without preoperative ERCP.

Instead of ERCP, the surgeon should be encouraged to perform intraoperative cholangiography (IOC). If that procedure detects one or more stones in the bile duct, ERCP can then be performed the following day or as soon as possible. Patients who have persistent or progressive biliary obstruction, with or without radiologic evidence of choledocholithiasis, should undergo ERCP preoperatively to remove the stones that are almost always present. Once the bile duct is cleared by the endoscopist, surgery should follow promptly to prevent further migration of stones from the gallbladder. Although endoscopic opening of the biliary sphincter should prevent further stone impaction, that procedure does not always succeed, as most of us who perform such procedures have learned by experience.

Some surgeons demand that preoperative ERCP be performed because they believe that IOC unduly prolongs laparoscopic cholecystectomy or because they simply "don't do" IOC. The ability to perform IOC is part of the necessary skill set of surgeons performing laparoscopic surgery; endoscopists should not be performing routine endoscopic cholangiography as an alternative. Because ERCP and sphincterotomy combined are associated with much higher morbidity and mortality than is laparoscopic cholecystectomy—partly owing to the large number (more than 500,000) of cholecystectomy procedures performed annually in the U.S. alone—endoscopists performing ERCP prior to laparoscopic cholecystectomy must consider the medicolegal consequences in the event of a severe complication (usually severe pancreatitis) related to the procedure.

A surgical argument in favor of preoperative ERCP in resolving gallstone pancreatitis is that the surgeon must know that the endoscopist can access the duodenal papilla, perform cholangiography, and, if necessary, remove the offending stone or stones before surgery, so that an alternative treatment strategy can be planned if the operation were to fail. In expert hands, the success rate of ERCP, sphincterotomy, and stone removal should be in the range of 90% to 95%. Unfortunately, the success rate of cholangiography as performed by the many inexperienced and undertrained endoscopists in the community is 50% or less. Worse still, these endoscopists often perform diagnostic cholangiography but are unable or unwilling to attempt biliary stone removal.

Those of us who practice at referral centers often receive patients who must undergo urgent repeat ERCP to remove the retained biliary stones. Typically, the endoscopists who referred them ask that the patients be transferred back to them so that those patients can undergo a cholecystectomy performed by the referring endoscopists' surgical colleagues. Endoscopists who perform diagnostic ERCP but cannot undertake basic therapeutic procedures, such as stent placement and biliary sphincterotomy, have no place in modern endoscopic practice.

An exception to the "straight to surgery" recommendation for patients who have uncomplicated, recovered gallstone pancreatitis is the presence of suspected or known surgical reconstruction of the upper gut, such as Billroth-II partial gastrectomy and Roux-en-Y biliary diversion. Because the success rate of ERCP is relatively low after surgery, even in expert hands, preoperative ERCP may be appropriate to guide the surgeon's subsequent approach.

The problem of highly variable ERCP expertise around the country can be resolved only through selective training, so that fewer endoscopists and only those with greater experience and expertise perform this procedure. Increasingly, ERCP is being performed at local and regional referral centers, where biliary and pancreatic disorders are managed by a multidisciplinary team.

9. Should all patients with acute pancreatitis receive prophylactic antibiotic therapy?

No. Prophylactic antibiotic therapy is indicated only for patients with predicted severe acute pancreatitis. The antibiotic of choice is imipenem, the only one shown in a prospective, randomized clinical trial to reduce morbidity and mortality when pancreatic necrosis is present. A low-grade fever is common in acute pancreatitis and reflects the inflammatory process. A high temperature (greater than 38.5°C) associated with a high white blood cell count is usually due to bacterial infection. The source should be sought in the cultures of blood, fluid from percutaneous drains and, when appropriate, needle aspirates from the pancreas itself.

10. Is the finding of "infected necrosis" in acute pancreatitis really an indication for immediate surgical debridement or necrosectomy?

No. This question is being debated among researchers, but most experts would agree that a failure to respond to broad-spectrum parenteral antibiotic therapy is the indication for urgent necrosectomy. The procedure is not necessary for a patient whose fever and elevated white blood count resolve after antibiotic therapy and whose collected fluid can be removed via percutaneous drainage. However, about half of all patients with 50% or more of the pancreas involved in necrosis will eventually require surgery.

11. How should I approach patients who return with repeated attacks of pancreatitis but have already had their gallbladder removed?

Patients who have recurrent or relapsing pancreatitis—that is, serial attacks of acute pancreatitis—after undergoing cholecystectomy may have a variety of underlying disorders, ranging from retained biliary calculi, sphincter of Oddi dysfunction, and pancreas divisum, to unsuspected hyperparathyroidism, genetic (hereditary) pancreatitis, and pancreatic cancer. The evaluation of these disorders is beyond the scope of this discussion, but patients with idiopathic pancreatitis are best referred to a specialist center where expert ERCP, including sphincter of Oddi and pancreatic manometry and bile crystal analysis, is available.