Acute and Necrotizing Pancreatitis
One hundred and twenty years ago Reginald Fitz, a Harvard anatomist, systematically described a fatal case of acute pancreatitis and recommended against surgical intervention. In 1929 the advent of amylase assay allowed the identification of early pancreatitis, and through the 1930s and 1940s the prevailing trend was against surgical intervention. In 1959 Pollack advocated surgery for salvage of severe pancreatitis. This approach was adopted, but mortality remained high ( > 60%). The advent of contrast-enhanced CT scanning in 1985 allowed identification of necrotic, non-perfused/non-viable gland, and the extent of necrosis correlated with survival, and helped identify those patients at high risk who would need more intensive care. The majority (80%) of patients with acute pancreatitis have a mild/interstitial form that resolves with treatment of the underlying cause. About 20% of patients will undergo pancreatic necrosis and progress to the severe form of the disease. Of the latter 20%, about 40% will develop infected pancreatic necrosis, which is a major determinant of mortality. Increasing understanding of the pathophysiology, grading algorithms for early identification of high-risk patients, and advances in diagnostic imaging and intervention options are improving management of and recovery from severe pancreatitis.
Acute pancreatitis accounts for over 200,000 hospital admissions in the United States per year. The leading cause of pancreatitis is biliary stone disease (34-54%) followed by alcohol abuse. The biliary etiology most commonly involves small ( < 5 mm) stones or microlithiasis/sludge. A small percent of ethanol abusers develop pancreatitis, and ethanol is the most common cause of recurrent and chronic pancreatitis. Other less common causes of pancreatitis include hypertriglyceridemia, hypercalcemia, medications, tumor obstructing the pancreatic duct, ERCP, developmental anomalies (pancreatic divisum [7-10%], annular pancreas), and sphincter of Oddi dysfunction (~43% of idiopathic pancreatitis).
Premature activation of trypsin (conversion of trypsinogen to trypsin) within pancreatic acinar cells by one of a variety of proposed mechanisms triggers activation of several destructive pancreatic digestive enzymes and leads to cytokine-mediated intrapancreatic inflammation involving activated inflammatory cells, chemoattraction of those cells to the pancreatic microcirculation, activation of adhesion molecules which bind the inflammatory cells to the microcirculation, and migration of the inflammatory cells into areas of inflammation. The cascade of enzyme activation leads to autodigestion and damage to the pancreatic microcirculation, release of pancreatic enzymes into the bloodstream, and may trigger the systemic activation of inflammatory cytokines (systemic inflammatory response syndrome/SIRS). SIRS is defined as two or more of the following: P > 90, R > 20 or pCO2 < 32 mm, rectal T < 36 or > 38C, WBC < 4 or > 12. The inflammatory cytokines in turn can damage normal organs (kidneys, lungs) resulting in multi-organ failure syndrome and high (47%) mortality. The local pancreatic tissue damage predominates in the first 24-48 hours of an acute attack, and in the majority (2/3) of patients remains stable after that through the subsequent course of the disease. Complications include local, retroperitoneal, intraperitoneal and pleural fluid collections, infection/abscess of the necrotic gland, disruption of the pancreatic duct, psuedoaneurysm and bowel necrosis. Alcohol is a susceptibility factor that acts as an immune modulator that suppresses the innate proinflammatory response in favor of the anti-inflammatory response. Smoking is also a strong anti-inflammatory cytokine promoter.
The diagnosis of acute pancreatitis requires two of three criteria: characteristic pain (epigastric penetrating to back), amylase/lipase elevation to or greater than 3X upper limit of normal, or characteristic CT findings. The Atlanta Symposium on Severe Acute Pancreatitis of 1992 defined acute pancreatitis as an inflammatory process of the pancreas that may also involve peripancreatic tissue and/or remote organ systems.
Differential diagnosis includes mesenteric ischemia/infarct, perforated ulcer, biliary colic, dissecting abdominal aortic aneurysm, intestinal obstruction and inferior wall MI. History of gallbladder disease, alcohol abuse, medications, trauma, malignancy and family history of pancreatitis should be elicited.
Early clinical signs of severe pancreatitis (flank ecchymosis/Gray-Turner sign; periumbilical ecchymosis/Cullen's sign) are rare ( < 3%), and CT findings are most evident only after 72 hours of evolution, so early risk stratification systems (Ranson, Apache II) have been formulated to identify patients early who are at high-risk for severe disease. Ranson in 1974 proposed a dual set of criteria, one on admission (age > 55, WBC > 16K, glucose > 200mg/dl, LDH > 350IU/L, AST > 250IU/L), and the second at 48 hours after onset (Hct drop > 10%, BUN > 8mg/dl, Ca++ < 8 mg/dl, arterial PO2 < 60 mm Hg, base deficit > 4meq/L and fluid requirement > 6L). Assigning one point for each of the abnormalities, the predicted mortality for a score < 3 is 1, 3-4 = 15%, 5-6 = 40% and > 6 = 100%. Like CT, early prediction before 48 hours using Ranson criteria is problematic.
More recently, the Acute Physiology and Chronic Health Evaluation II (APACHE II) score measuring 12 routine physiological parameters (vital signs, oxygenation, blood chemistries, Hct and WBC count) has been proposed, and has been found to be at least equivalent to Ranson criteria in predictive value, and has the advantage of being applicable in the first 48 hours. Additional points can be added to the APACHE II score for age and chronic health problems to try to refine the predictive value. There are on-line calculators for entering APACHE data and calculating score. Other systems such as the Glasgow criteria use similar parameters. Disseminated intravascular coagulation (DIC: pl < 100 k/mm3, fibrinogen < 100 mg/dL, fibrin split products > 80 micrograms/ml, Ca++ < 7.5 mg/dL) may also be predictive. Obese patients (BMI > 30Kg/m2) have more local and systemic complications, but not a higher mortality. After 72 hours, elevated ( > 150mg/L) C-reactive protein, an acute phase reactant, is proportional to the degree of pancreatic necrosis. Earlier chemical markers of severity are under investigation, but not yet clinically applicable.
The Atlanta Symposium included a Ranson score of 3 or above, an APACHE II score of 8 or above, organ failure and local complications in their definition of severe pancreatitis. The organ failure criteria included shock (systolic < 90 mm), hypoxia (PaO2, 60mm Hg), and creatinine > 2 mg.L after hydration. GI bleeding will likely be deleted from the criteria.
The imaging modality of choice for pancreatitis is CT. Early ( < 48-72h) in the course of the disease, the role of CT (without IV contrast) is in differential diagnosis. After that time, multi-detector, contrast enhanced (CECT) CT is the standard of care, giving high resolution with a short breath-hold. The benefit of identifying the presence and degree of necrosis outweighs the potential harmful effect of the IV contrast decreasing pancreatic tissue oxygenation. The single portal venous phase is usually sufficient, but arterial phase is included when vascular complications are suspected. Post-processing of 2D and 3D images is essential. Recently quantitative pancreatic blood flow measurement has been added. Early fluid collections are seen in 30-50% of acute pancreatitis. Half of these resolve without intervention. The remainder may progress to pseudocyst formation, usually after 4 weeks.
The Balthazar classification of acute pancreatitis is a scale from A-E: A = normal, B = focal or diffuse enlargement, C = gland and mild peripancreatic changes, D = single fluid collection, E = 2 or more fluid collections or gas. Class D or E carries a 14% mortality and 54% morbidity.
Pancreatic necrosis is indicated by non-enhancing pancreatic parenchyma, which becomes evident after about 72 hours. It may be focal or diffuse. Pancreatic attenuation less than 30 Hounsfeld units with CECT indicates hypoperfusion. Fatty infiltration or edema may give a false positive reading. The extent of necrosis is directly correlated with mortality and morbidity, with a dramatic increase over 30% necrosis.
CT also identifies local complications, including solid organ involvement (spleen, liver, kidney) and bowel (late). Finally, percutaneous, CT-guided interventions are an increasingly useful tool in the management of complicated pancreatitis.
MRI is useful in specific situations. It is an alternative if a patient can't tolerate iodinated contrast media, in a young patient to avoid the radiation exposure of multiple CT scans and to characterize the consistency of fluid collections for drainage better than CT. MR angiogram can be used for suspected vascular complications, and MRCP is useful for defining the pancreatic ducts (duct disruption, communication with pseudocyst. It is also capable of identifying pancreatic anomalies (annular, divisum) and neoplasm. Three-D MRCP has recently become available. Administration of secretin can distend the pancreatic duct and enhance visualization depending on the presence of functional pancreatic tissue.
Ultrasound has been the initial procedure of choice to assess the role of biliary stone disease as the etiology and to identify patients in need of ERCP intervention. The study is limited by the presence of ileus which often accompanies acute pancreatitis. Compression and change of position may help visualization. Size of the gland, clarity of margins, homogeneity and focal changes may be assessed. Endoscopic ultrasound for common duct stones has been supplanted by MRCP. Ultrasound technology is improving, and ultrasound may also help diagnose necrosis when IV bubble agents become available in the US.
The goal of initial management is to prevent decrease of oxygen delivery to the pancreas. Capillary leak and peripancreatic inflammation lead to third space fluid loss and hypovolemia. The decreased organ perfusion which can contribute to further organ necrosis (initial necrosis begins early, before large volume shifts) is counteracted by vigorous crystalloid resuscitation. Volume is also critical to prevent intestinal ischemia and translocation of intestinal enteric organisms which increase the risk of infected pancreatic necrosis.
Close monitoring of physical exam, vital signs, urine output, acid-base balance and oxygenation is necessary, as clinical signs are the most reliable indication of course. Supplemental oxygen should be administered in the first 24-48 hours if there is suspicion of severe pancreatitis, especially if oxygen saturation falls below 95%.
Excessive fluid requirements, organ dysfunction or comorbidities may indicate early transfer to a step-down unit or ICU. Patients with cardiac dysfunction may present problems assessing and titrating fluid requirements and may need Swan-Ganz monitoring.
NG tube is indicated if there is gastric distention. Intravenous narcotics are administered as well as prophylactic proton pump inhibitor (stress and decreased bicarbonate secretion predisposing to ulcer) and DVT prophylaxis.
Mild interstitial pancreatitis may allow early feeding with resolution, most within 5-7 days, but necrotizing pancreatitis may require initial NPO. Prior management of keeping patients with severe pancreatitis NPO has been supplanted by earlier initiation of enteral feeding. Enteral feeding by nasojejunal or nasogastric tube is preferred over TPN for those patients requiring long term alimentation because of its effect on stabilizing the gut barrier and decreasing bacterial translocation, decreasing the immune response, and the incidence of SIRS. Most literature supports semielemental tube feeding. Intolerance to the tube feeds in the form of bloating and increased pain may necessitate TPN.
Targeted treatments of the inflammatory cascade such as octreotide have not proven efficacious. Prophylactic antibiotics are not recommended except in a symptomatic patient while investigating the source of the sepsis.
For the majority of patients with biliary pancreatitis, the course is mild, and the gallbladder should be removed after the pancreatitis subsides after 48 hours. Failure to remove the gallbladder results in 80% recurrence of pancreatitis within one year. Early ERCP and sphincterotomy is not indicated in mild biliary pancreatitis unless there is persistent biliary obstruction or cholangitis. Sphincterotomy may be indicated for sphincter of Oddi dysfunction, and of the narrow dorsal duct orifice in pancreatic divisum. There is a 10-20% restenosis rate and initial stent placement is recommended.
The incidence of infected pancreatic necrosis, historically 70%, now runs 10-40%, with the organisms most commonly coming from translocated enteric flora. Antibiotic use has changed the pattern of infection from Gram negative (e.coli most commonly) to Gm+ (enterococcus, staph). Infection increases the predicted mortality by 3X. Therapeutic antibiotics in the face of clinically suspected or proven infection should penetrate pancreatic tissue and cover the expected flora. These include the carbapenems, quinalones and nitroimidazoles.
The degree of necrosis is proportional to mortality, greater than 50% increasing mortality 8X. Sterile necrosis is best treated non-operatively, with a 2-5% mortality. Within 2-3 weeks a capsule forms around the pancreatic necrosis (organized pancreatic necrosis). Surgery for sterile necrosis increases mortality to 12-21%. Some patients with sterile necrosis may ultimately require surgery, e.g. refeeding pancreatitis, disconnected duct syndrome.
Abdominal compartment syndrome (intraabdominal pressure > 25 mm Hg) as a complication of pancreatic necrosis may lead to multiorgan failure and increased mortality, and necessitate decompression.
Infected pancreatic necrosis
Infected pancreatic necrosis necessitates debridement and drainage (100% mortality without). The traditional open surgical drainage techniques have now been supplemented with retroperitoneal, laparoscopic, CT-guided interventional, and endoscopic modalities. When infection supervenes, it is usually after 10 days, and 48% of patients with infected pancreatic necrosis have persistent organ failure. Leukocytosis, toxicity and organ failure do not differentiate sterile from infected necrosis. Therefore, most authorities recommend CT-guided fine needle aspiration (FNA) for Gram stain and culture to document infection and guide antibiotic coverage. There is discussion whether surgical debridement should be done on documenting infection, or after a course of antibiotics.
Surgical necrosectomy strategies, closed with continuous drainage, with or without irrigation, have proven equivalent. Open technique usually involves a subcostal incision and entry into the lesser sac. If timing was appropriate (surgery should be as late as possible) to allow the pancreatic necrosis to demarcate, the loose pancreatic tissue is gently removed with blunt finger dissection. Formal resection carries a high morbidity. The retroperitoneum, paracolic gutters and mesentery are gently explored. If gallstones are the etiology, and if feasible, the gallbladder is removed. A feeding jejunostomy can be placed, but there is a high incidence of enterocutaneous fistula. Several large-bore closed suction drains are placed. An alternative, less invasive retroperitoneal approach has recently been advocated, leaving a wide drainage tract.
Laparoscopic treatment of infected pancreatic necrosis is evolving. Small numbers of selected patients have been treated in this way, with size and location being important factors. Laparoscopic debridement and drainage may be indicated for patients with major comorbidities as a temporizing or definitive treatment.
Interventional drainage techniques are primarily indicated for large collections, and may require placement of multiple large bore catheters. Percutaneous drainage of chronic pseudocysts is also indicated.
Endoscopic transgastric/transduodenal drainage (VID 1163) for favorable collections has been tried with significant (45%) serious complications and later need for surgical debridement and drainage.
Interventional radiology embolization is indicated for pseudoaneurysm and hemorrhage (VID 151) into a pseudocyst.
Banks A et al; Practice Guidelines in Acute Pancreatitis; Am J Gastroenterology 2006; 101:2379-2400
Rau B; Outcome determinants in acute pancreatitis; Am J Surg 194 (suppl to Oct 2007) S39-S44
Bradley E L III; Surgery of acute pancreatitis; Am J Surg 194 (suppl to Oct 2007) S20-S23
Saokar A et al; Cross-sectional imaging in acute pancreatitis; Radiol Clin N Amer 45 (2007) 447-460
Siriboon A et al; Endoscopic management of acute and chronic pancreatitis; Sur Clin N Amer 87 (2007) 1379-1402
Yadav D et al; Alcohol-associated pancreatitis; Gastroenterol Clin N Amer 36 (2007) 219-238
Tchelepi H; Ultrasound of acute pancreatitis; Ultrasound Clin 2 (2007) 415-422
Werner J et al; The role of minimally invasive techniques for necrosectomy in acute pancreatitis; Am J Surg (Suppl to Oct 2007) S24-S27
Curtis C, Kudsk K; Nutrition support in pancreatitis; Surg Clin N Am 87 (2007) 1403-1405
Haney J, Pappas T; Necrotizing pancreatitis: diagnosis and management; Surg Clin N Am 87 (2007) 1431-1446
Harrison D et al; The pancreatitis outcome prediction (POP) score: a new prognostic index for patients with severe acute pancreatitis; Crit Care Med 35:7 (2007) 1703-1708