Figure 53-1. Schematic diagram depicting activation of proteolytic enzymes, possibly through colocalization of zymogen granules and lysosomes, and subsequent rupture of zymogen granules releasing the activated enzymes into the cytoplasm of the pancreatic acinar cell. The activated enzymes then undergo disordered basolateral discharge from the acinar cell into the pancreatic parenchyma.
Acinar cell injury induced by active trypsin allows it to be released into the pancreatic parenchyma (Fig. 53-1) where it activates more trypsin and other digestive enzymes (e.g., chymotrypsin, phospholipase, and elastase). Trypsin can also activate the complement, kallikrein-kinin, coagulation, and fibrinolysis cascades within the pancreatic parenchyma. Activation of these enzymes is believed to initiate a vicious cycle in which activated enzymes cause cellular injury, an event that leads to the release of even more destructive enzymes. This cycle can overwhelm defense mechanisms that normally serve to limit the injurious consequences of premature trypsin activation within the pancreas (e.g., pancreatic secretory trypsin inhibitor–mediated inhibition of trypsin activity).
An inflammatory response is then generated in response to the initial acinar cell injury. This inflammatory response is marked by the infiltration of the pancreatic parenchyma with immune cells such as neutrophils, macrophages, monocytes, and lymphocytes and the release of a broad range of proinflammatory mediators such as tumor necrosis factor (TNF) α; interleukins (IL) 1β, 6, and 8; platelet-activating factor; chemokines (i.e., CXCL2 and CCL2); prostaglandins; and leukotrienes. The inflammatory response, to a large extent, determines the severity of pancreatitis, and the blockade of several components of the inflammatory response ameliorates the disease and reduces mortality in experimental models. The understanding of how the initial acinar cell injury provokes an inflammatory response is incomplete, but it appears that reactive oxygen species (ROS) and innate molecular pattern recognition (i.e., damage/danger-associated pattern molecules and toll-like receptors (TLRs) as well as the activation of transcription factors such as NF-κB play a role.8
The result of these events is pancreatic autodigestion, with injury to the vascular endothelium, interstitium, and acinar cells. Increases in vascular permeability lead to interstitial edema. Vasoconstriction, thrombosis, and capillary stasis can lead to ischemic (and perhaps ischemia–reperfusion) injury and the development of pancreatic necrosis. With severe pancreatic injury, the systemic inflammatory response syndrome (SIRS) and distant organ failure can occur. The systemic complications are believed to be mediated by digestive enzymes and inflammatory mediators released from the injured pancreas. For example, activated phospholipase A–induced digestion of lecithin (an important component of pulmonary surfactant) may play a role in pathogenesis of acute respiratory distress syndrome (ARDS) that occurs in the setting of acute pancreatitis. In addition, the circulatory and inflammatory effects induced by acute pancreatitis are postulated to impair intestinal epithelial barrier function, allowing for the translocation of bacteria from the intestinal lumen into the systemic circulation. This phenomenon has been demonstrated to occur in animal models and may account for the pathogenesis of pancreatic and peripancreatic infection that can complicate necrotizing pancreatitis.
Gallstones cause approximately 35% of episodes of acute pancreatitis in the United States. In a mechanistic model proposed over a century ago, a gallstone lodged at the papilla of Vater occludes the ampullary orifice, leading to retrograde reflux of bile into the pancreatic duct through a common channel shared by the common bile duct and the pancreatic duct (Fig. 53-2). Although elements of this model have been challenged, the prevailing view is that transient or persistent obstruction of the ampullary orifice by a gallstone or edema induced by stone passage is the inciting factor in the pathogenesis of gallstone-induced pancreatitis. Microlithiasis refers to aggregates (<5 mm in diameter) of cholesterol monohydrate crystals or calcium bilirubinate granules detected as “sludge” within the gallbladder on ultrasonography or on examination of bile obtained during endoscopic retrograde cholangiopancreatography (ERCP). An etiologic role for microlithiasis in acute pancreatitis remains unproved; however, data derived from case-control studies suggest that cholecystectomy or endoscopic sphincterotomy can reduce the risk of recurrent acute pancreatitis in patients with microlithiasis.
Figure 53-2. Illustration of the common channel concept. A gallstone lodged at the ampulla of Vater can cause reflux of bile into the pancreatic duct.
Ethanol causes approximately 40% of cases of acute pancreatitis in the United States. Most patients with alcohol-induced acute pancreatitis also have underlying chronic pancreatitis. Potential mechanisms by which alcohol-induced pancreatitis include sphincter of Oddi spasm, obstruction of small pancreatic ductules by proteinaceous plugs, alcohol-induced metabolic abnormalities (e.g., hyperlipidemia), and direct toxic effects induced by alcohol and its metabolites (e.g., acetaldehyde, acetate, and nonesterified fatty acids).
A wide range of other etiologies of acute pancreatitis have been identified (Table 53-2). Ongoing investigations are beginning to reveal specific gene abnormalities (e.g., mutations in cationic trypsinogen PRSS1, pancreatic secretory trypsin inhibitor SPINK1, and the cystic fibrosis transmembrane conductance regulator CFTR) that can be associated with pancreatitis. Patients for whom no etiology can be identified despite thorough evaluation are classified as having idiopathic pancreatitis.
Table 53-2 Etiology of Acute Pancreatitis
Abdominal pain, nausea, and vomiting are the most prevalent symptoms associated with acute pancreatitis. The pain is visceral in quality, is localized to the epigastrium, often radiates to the back, and may be alleviated with the patient leaning forward. Abdominal tenderness is the most prevalent sign of acute pancreatitis. Tachycardia and hypotension may result from intravascular hypovolemia. Low-grade fevers are common, but high-grade fevers are unusual in the absence of intra- or extrapancreatic infection. Jaundice may be evident in the presence of cholangitis (e.g., with gallstone-induced pancreatitis and persistent choledocholithiasis) or liver disease (alcohol-induced pancreatitis in a patient with cirrhosis). Evidence of retroperitoneal hemorrhage may be become apparent if blood dissects into the subcutaneous tissues of the flanks (Grey Turner sign), umbilicus (Cullen sign), or inguinal region (Fox sign); however, these findings are unusual. In approximately 20% of cases, acute pancreatitis is associated with SIRS, hemodynamic lability, and/or organ failure (particularly compromise of the cardiovascular, pulmonary, and renal systems) on presentation.
With pancreatic injury, a variety of digestive enzymes escape from acinar cells and enter the systemic circulation. Of these enzymes, amylase is the most widely assayed to confirm the diagnosis of acute pancreatitis. Amylase levels rise within several hours after onset of symptoms and typically remain elevated for 3 to 5 days during uncomplicated episodes of mild acute pancreatitis. Because of the short serum half-life of amylase (10 hours), levels can normalize as soon as 24 hours after disease onset. The sensitivity of this test depends on what threshold value is used to define a positive result (90% sensitivity with a threshold value just above the normal range vs. 60% sensitivity with a threshold value at three times the upper limit of normal). Specificity (which also varies with the threshold values selected) is limited because a wide range of disorders can cause elevations in serum amylase concentration. Assays that detect increases in the serum concentration of amylase of pancreatic origin (P-isoamylase) alone are associated with greater specificity. Increased urinary amylase concentrations and amylase-to-creatinine clearance ratios occur with acute pancreatitis; however, these parameters offer no advantage over serum amylase concentrations, except in the evaluation of macroamylasemia (in which urinary amylase excretion is not increased despite elevations in serum amylase concentration).
Serum lipase concentrations increase with kinetics similar to those of amylase. It has a longer serum half-life than amylase, however, and may be useful for diagnosing acute pancreatitis late in the course of an episode (at which time serum amylase concentrations may have already normalized). Although lipase is more specific than amylase in the diagnosis of acute pancreatitis, note that lipase is produced at a range of nonpancreatic sites, including the intestine, liver, biliary tract, and stomach, and tongue.
The magnitude of the increases in amylase or lipase concentrations has no correlation with severity of pancreatitis. In general, the magnitude of increases in amylase concentrations tends to be greater in patients with gallstone pancreatitis than in those with alcohol-induced pancreatitis; however, this finding is unreliable in distinguishing between these two etiologies.
Findings on plain radiographs associated with acute pancreatitis are nonspecific and include ileus that may be generalized or localized to a segment of small intestine (“sentinel loop”) or transverse colon (“colon cut-off sign”), psoas muscle margins that are obscured by retroperitoneal edema, an elevated hemidiaphragm, pleural effusions, and basilar atelectasis.
Ultrasonography may reveal a diffusely enlarged, hypoechoic pancreas. However, overlying bowel gas (particularly prominent with ileus) limits visualization of the pancreas in a large percentage of cases. Although ultrasonography has poor sensitivity in the diagnosis of acute pancreatitis, it plays an important role in the identification of the etiology of pancreatitis (e.g., the detection of gallstones).
CT scanning is the most important imaging test in the evaluation of acute pancreatitis. CT findings of mild acute pancreatitis include pancreatic enlargement and edema, effacement of the normal lobulated contour of the pancreas, and stranding of peripancreatic fat (Fig. 53-3). In addition, dynamic CT scanning performed after the bolus administration of intravenous contrast can demonstrate regions of pancreas that have poor or no perfusion, as seen with pancreatic necrosis (Fig. 53-4). Detection of necrosis plays an important role in assessment of disease severity, as discussed further later. CT can also characterize collections and other complications associated with acute pancreatitis.
Figure 53-3. Computed tomography scan of acute interstitial pancreatitis.
Figure 53-4. Computed tomography scan of acute necrotizing pancreatitis.
MRI and magnetic resonance cholangiopancreatography (MRCP) are being used with increasing frequency in patients with acute pancreatitis. These examinations have the potential to offer better definition of pancreatic and biliary ductal abnormalities than CT scanning, and they are applicable in patients for whom ionizing radiation or iodinated intravenous contrast agents used in CT scanning are contraindicated. MRI can suggest the presence of pancreatic necrosis even without the use of intravenous gadolinium. MRI also has the advantage of better characterizing collections associated with acute pancreatitis, in particular with respect to differentiating solid from liquid components. Disadvantages of MRI include high cost, limited availability, and the long duration of examinations.
ASSESSMENT OF DISEASE SEVERITY
Accurate prediction of severity early in the course of disease offers potential benefits in that complications can be anticipated and detected early through the use of intensive monitoring and frequent clinical assessment, and early and aggressive therapies can be instituted to attempt to prevent these complications. Several methods for assessing disease severity based on clinical parameters, serum markers, imaging, and scoring systems have been widely studied.
Revised Atlanta and Determinant-Based Classifications
An optimal system for the classification of severity of acute pancreatitis has not been universally accepted. Nevertheless, it is widely recognized that organ failure and local complications, particularly necrosis, are the two most important predictors of disease severity. While these two variables are often associated, they differ as to when in the course of disease they typically arise and the manner in which they contribute to disease severity. The early phase of acute pancreatitis (generally the first week after onset of symptoms) is characterized by the host response to the local pancreatic injury. Disease severity in the early phase is thus determined by organ failure resulting from the systemic inflammatory response. Organ failure can be immediately apparent on clinical presentation, and in fact, organ failure occurring in the early phase of disease is usually present on admission. In contrast, pancreatic necrosis, which occurs in 5% to 10% of cases of acute pancreatitis, may not be apparent on initial presentation and the extent of local complications can continue to evolve into the late phase of disease. Organ failure and death in the late phase are often determined by local complications, particularly infected necrosis.3 While local complications can contribute to the mortality of patients with organ failure, there is also a subset of patients with pancreatic necrosis without organ failure who experience significant morbidity but have very low mortality.13
Two classification schemes based on organ failure and local complications have been proposed (Table 53-3).3,13,15 While these classification schemes are quite useful and validated in predicting mortality,11,14 their utility in predicting the course of disease on admission or within the first few days of hospitalization is limited by the fact that local complications and the resultant organ failure occurring in the late phase of the disease have not become apparent. Furthermore, a distinction is made between transient organ failure and persistent organ failure. This distinction cannot be made at the time of presentation. Similarly, the distinction between sterile necrosis and infected necrosis in the Determinant-Based Classification cannot be made with certainty until a patient with necrotizing pancreatitis shows signs of resolution.
The Ranson criteria, which is based on age, white blood cell count (WBC), glucose, serum lactate dehydrogenase (LDH), and serum aspartate aminotransferase (AST) all determined on admission as well as hematocrit drop, blood urea nitrogen (BUN), serum calcium, arterial partial pressure of oxygen (PaO2), base deficit, and fluid requirement all determined after 48 hours, are easily tabulated, and the resulting scores are well correlated with morbidity and mortality rates.16 The presence of three or more of these criteria is indicative of severe acute pancreatitis. Important limitations of the Ranson criteria are that the predictive score cannot be determined prior to 48 hours following admission and that it can only be used once. Furthermore, because these criteria were developed using a cohort of patients for whom alcohol was the predominant etiology of pancreatitis, their generalizability may be limited. A similar predictive scoring system developed in Glasgow using a cohort of patients for whom gallstones were the predominant etiology of pancreatitis is available.17
Acute Physiology and Chronic Health Evaluation (APACHE) II scores, which are based on patient age, indices of chronic health, and physiologic parameters, can be determined at any time after admission, can be updated continuously during the course of disease and may have greater predictive power than Ranson scores.18 However, the complexity of calculating APACHE II (or related APACHE III) scores limits its application in routine clinical practice.
Recently another scoring system, the Bedside Index for Severity in Acute Pancreatitis (BISAP), based on five relatively straightforward parameters that are obtained within the first 24 hours of hospital admission, has been proposed. One point is assigned for the presence of each of the following: BUN >25 mg/dL, impaired mental status, SIRS, age >60 years, and the presence of a pleural effusion.19 The BISAP has been validated and is comparable to the Ranson criteria20,21 and the APACHE II Score19,21,22 in its prediction of mortality.
Computed Tomography Scanning
The diagnostic application of CT scanning was discussed previously. Although not required for making a diagnosis on admission, a CT is often useful in patients who deteriorate or fail to improve. This technique is associated with greater than 90% sensitivity in the detection of pancreatic necrosis, a finding that is predictive of disease severity. Furthermore, CT scanning can also diagnose and characterize collections and can indicate infection (e.g., if air bubbles are present) in some cases. Because necrosis takes time to develop (in some patients up to 5 days), a contrast-enhanced CT scan obtained too early in the disease course (e.g., at the time of admission) does not have predictive value beyond clinical assessment (i.e., APACHE II or BISAP scoring).23 In addition, concerns that early administration of iodinated intravenous contrast agents used in CT scanning may exacerbate pancreatic injury have been raised, although these agents have not been shown to cause or exacerbate pancreatitis in humans.24
Serum and Urinary Markers
C-reactive protein (CRP) is an easily assayed marker for which serum concentrations are well correlated with disease severity. However, CRP levels do not become significantly elevated until 48 hours after onset of disease; therefore, this marker is not useful for early prediction of disease severity. Lipopolysaccharide-binding protein (LBP), a class 1 acute-phase protein that binds and transfers bacterial lipopolysaccharide (LPS), has also been shown to correlate with disease severity but its elevation is similarly delayed. Serum concentrations of IL-6, IL-8, neutrophil elastase, angiopoietin-2, procalcitonin, and urinary concentrations of TAP (a product of trypsinogen activation) are also correlated with disease severity. Because these markers become elevated within 24 hours of disease onset, they may be relevant to early prediction strategies in the future. Currently, however, assays for these markers are not widely available.
The goals of management for patients with acute pancreatitis are summarized below.
Correcting Pathophysiologic Derangements and Ameliorating Symptoms
Patients with potentially severe pancreatitis, as well as those for whom initial resuscitation fails, are best managed in a dedicated intensive care unit (ICU). Central venous monitoring may facilitate fluid management. Patients should be closely monitored for development of distant organ failure, particularly respiratory, cardiovascular, and renal failure, so that supportive management of these conditions (positive-pressure ventilation, administration of vasopressor agents, and hemodialysis, respectively) can be instituted without delay.
Abdominal pain is usually ameliorated with intravenous narcotics. In the past, morphine was avoided because of concerns that morphine-induced increases in sphincter of Oddi pressure might exacerbate an episode of pancreatitis. However, there is no clinical evidence that morphine can induce or exacerbate acute pancreatitis. Other analgesic agents commonly used in patients with acute pancreatitis include meperidine and fentanyl. Evacuation of gastric contents using a nasogastric tube should be instituted if vomiting is a prominent symptom; otherwise, it is unnecessary.
A fraction of patients with severe pancreatitis develop abdominal compartment syndrome (ACS). ACS is defined as a life-threatening sustained elevation of intra-abdominal pressure associated with new-onset organ failure or acute worsening of existing organ failure. ACS usually manifests with a tensely dilated abdomen, oliguria, and increased peak airway pressures. The diagnosis and treatment of ACS is discussed in more detail elsewhere in this book. In acute pancreatitis, ACS is associated with very severe disease, and approximately half of patients with ACS do not survive. ACS is treated by attempting to decrease intra-abdominal pressure. In some patients this can be accomplished by draining intra-abdominal collections and/or decompressing the bowel with nasogastric and rectal tubes. More than half of the patients with ACS in the setting of acute pancreatitis require a decompressive laparotomy.9,26
Minimizing Progression of Pancreatic Inflammation and Injury
Identification of strategies for interrupting the inflammatory cascades that induce pancreatic injury and distant organ failure is an area of active investigation. Currently, the only method used in routine clinical practice is bowel rest (nothing by mouth). The rationale underlying this approach is that avoidance of bolus oral nutrient intake may limit stimulation of pancreatic exocrine secretion induced by the presence of nutrients in the intestine, particularly the duodenum.
Patients with mild acute pancreatitis generally need no nutritional support, as their disease typically resolves within 1 week. In contrast, patients with severe pancreatitis usually have a more prolonged disease course and should begin to receive nutritional support as early as feasible. Although these patients traditionally have been administered total parenteral nutrition (TPN), accumulating evidence suggests that enteral nutrition is safe, is less costly, and may be associated with a lower complication rate than TPN.27 Administration of enteral nutrition is also associated with the theoretical advantage of helping to maintain the integrity of the intestinal mucosal barrier, thus potentially limiting or preventing bacterial translocation. Enteral nutrients have typically been delivered to the jejunum through nasojejunal tubes to avoid stimulating pancreatic exocrine secretion, however, randomized controlled trials indicate that continuous feedings delivered through nasogastric tubes are equally safe and effective.28,29 TPN is still required in many patients who do not tolerate enteral nutrition due to ileus.
Clinical trials of agents that inhibit activated pancreatic enzymes, inhibit pancreatic secretion, or interrupt the inflammatory cascade have yielded disappointing results. Meta-analyses of clinical trials of gabexate mesylate (a proteinase inhibitor), somatostatin, and octreotide suggest these agents have limited, if any, efficacy in improving outcomes in acute pancreatitis. A platelet-activating factor antagonist, lexipafant, showed promise in an initial study but not in a subsequent larger trial and is not currently recommended. Other adjuncts for which clinical trials have failed to demonstrate efficacy in limiting pancreatic injury in patients with acute pancreatitis include glucagon, anticholinergics, fresh frozen plasma, and peritoneal lavage.
Treating the Underlying Cause
There is an approximately 18% incidence of recurrent pancreatitis or other gallstone-related complications during the 6-week period following an episode of gallstone pancreatitis in patients who do not undergo cholecystectomy.31 Indeed, there is a substantial incidence of recurrent pancreatitis and gallstone-related complications in patients with biliary pancreatitis whose cholecystectomy is delayed even 2 weeks after hospital discharge.32 Therefore, cholecystectomy should be performed during the same hospitalization for most patients with mild gallstone pancreatitis. This strategy does not increase operative complications, conversion to open procedures, or mortality.31 In severe biliary pancreatitis, the risk of deferring surgery needs to be balanced against the risk of performing early surgery in patients who are debilitated and nutritionally compromised. Endoscopic sphincterotomy is another option, however, the high risk of recurrent gallstone-related complications and the higher mortality compared to cholecystectomy suggest that this strategy should be reserved for patients with severe comorbidities precluding safe surgery.31,33
Examples of other measures directed at correcting the underlying cause of pancreatitis include cessation of drugs known to cause pancreatitis and treatment of hypercalcemia or hyperlipidemia.
Preventing and Treating Complications
Complications of acute pancreatitis include pancreatic abscesses and infected necrosis, pseudocysts and walled-off necrosis, pancreatic ascites and fistulas, splenic vein thrombosis, and arterial pseudoaneurysms (Table 53-4). Infected necrosis as well as pseudocysts and sterile walled-off necrosis are discussed in detail below.
Table 53-4 Local Complications of Acute Pancreatitis