Correction. (originally cited in print as reference #219; now cited as references 195 and 222)

Transcription

1 GASTROENTEROLOGY 2007;133:1056 Correction Pandol SJ, Saluja AK, Imrie CW, Banks PA. Acute pancreatitis: bench to the bedside. Gastroenterology 2007;132: The published version of this review article, prepared in collaboration with four authors, accidentally contained several errors in the order of references. The corresponding author has organized the corrected reference list so that it corresponds with the placement of citations in the text and maintains the published total of 261 references. Thus, three of the corrected references numbers 144, 229, and 250 now contain multiple citations (e.g., 144a & 144b). They appear as follows: 144. Kubisch C, DiMagno MJ, Tietz AB, Welsh MJ, Ernst SA, Brandt-Nedelev B, Diebold J, Wagner AC, Goke B, Williams JA, Schafer C. Overexpression of heat shock protein Hsp27 protects against cerulein-induced pancreatitis. Gastroenterology 2004;127: (above reference previously omitted from the published text) 144a. Frossard JL, Pastor CM, Hadengue A. Effect of hyperthermia on NF-kappaB binding activity in cerulein-induced acute pancreatitis. Am J Physiol Gastrointest Liver Physiol 2001;280:G1157-G Kumar A, Singh N, Prakash S, Saraya A, Joshi YK. Early enteral nutrition in severe acute pancreatitis: a prospective randomized controlled trial comparing nasojejunal and nasogastric routes. J Clin Gastroenterol 2006;40: a. Berzin TM, Mortele KJ, Banks PA. The management of suspected pancreatic sepsis. Gastroenterol Clin North Am 2006;35: Connor S, Raraty MG, Howes N, Evans J, Ghaneh P, Sutton R, Neoptolemos JP. Surgery in the treatment of acute pancreatitis minimal access pancreatic necrosectomy. Scand J Surg 2005;94: a. Connor S, Ghaneh P, Raraty M, Sutton R, Rosso E, Garvey CJ, Hughes ML, Evans JC, Rowlands P, Neoptolemos JP. Minimally invasive retroperitoneal pancreatic necrosectomy. Dig Surg 2003;20: b. Connor S, Alexakis N, Raraty MG, Ghaneh P, Evans J, Hughes M, Garvey CJ, Sutton R, Neoptolemos JP. Early and late complications after pancreatic necrosectomy. Surgery 2005;137: Also, the following articles are now cited twice in the corrected reference list, again to preserve continuity: Johnson CD, Abu-Hilal M. Persistent organ failure during the first week as a marker of fatal outcome in acute pancreatitis. Gut 2004;53: (originally cited in print as reference #219; now cited as references 195 and 222) Buter A, Imrie CW, Carter CR, Evans S, McKay CJ. Dynamic nature of early organ dysfunction determines outcome in acute pancreatitis. Br J Surg 2002;89: (still cited as refrence # 261, also cited as reference 226 in the corrected list) The order of references has also completely changed in the corrected list, but the citations within the text of the article have remained the same. Please see the online version of this correction for the complete article, including the revised reference list (see supplementary material online at

2 GASTROENTEROLOGY 2007;133:1056.e e25 REVIEWS IN BASIC AND CLINICAL GASTROENTEROLOGY Acute Pancreatitis: Bench to the Bedside STEPHEN J. PANDOL,* ASHOK K. SALUJA, CLEMENT W. IMRIE, and PETER A. BANKS *Department of Medicine, Department of Veterans Affairs and University of California, Los Angeles, California; Department of Surgery, University of Minnesota, Minneapolis, Minnesota; Glasgow Royal Infirmary, Glasgow, Scotland, United Kingdom; and Department of Medicine, Division of Gastroenterology, Brigham and Women s Hospital, Boston, Massachusetts Acute pancreatitis is an acute inflammatory process of the pancreas that frequently involves peripancreatic tissues and at times remote organ systems. The severity of the disease varies widely from mild forms only affecting the pancreas to severe disease with multisystemic organ failure and death. In this review, a synopsis of our current understanding of the pathobiologic processes that underlie pancreatitis and its sequelae is provided; and a current approach to the management of patients with this disorder is reviewed. We conclude this review with suggestions for future directions for research to improve outcomes for patients who have acute pancreatitis. The major pathobiologic processes underlying acute pancreatitis are inflammation, edema, and necrosis of pancreatic tissue as well as inflammation and injury of extrapancreatic organs. The overall mortality in patients with acute pancreatitis is approximately 5%. 1 Mortality is higher in patients with necrotizing pancreatitis compared with those with interstitial pancreatitis in which there is little necrosis (approximately 17% vs 3%, respectively). Among patients with necrotizing pancreatitis, mortality is greater in patients with infected necrosis than in those with sterile necrosis (approximately 30% vs 12%, respectively). The prevalence of infected necrosis in patients with necrotizing pancreatitis now appears to be approximately 15% 20% compared with approximately 35% several years ago. Far more patients have interstitial pancreatitis than necrotizing pancreatitis (approximately 85% vs 15%, respectively). Organ failure occurs more commonly in patients with necrotizing pancreatitis compared with those with interstitial pancreatitis (approximately 50% vs 5% 10%, respectively). In necrotizing pancreatitis, mortality is very low in the absence of organ failure (close to 0%); in the presence of single organ failure, mortality is generally less than 10%; with multisystem organ failure, mortality is in the range of 35% 50%. Approximately half of deaths in acute pancreatitis occur within the first 2 weeks of illness and are generally attributed to organ failure. The remainder of deaths occurs weeks to months after this interval, and death is generally related to organ failure associated with infected necrosis or complications of sterile necrosis. Although alcohol abuse and gallstone disease account for 70% 80% of the cases of acute pancreatitis, the exact mechanisms by which these factors initiate acute pancreatitis are presently unknown. In addition, because of the rapid course of the disease and the relative inaccessibility of pancreatic tissue for examination during pancreatitis, investigations of the mechanisms underlying these pathobiologic processes have been hampered. Considering these obstacles, investigators have turned to animal models of acute pancreatitis to reveal the molecular steps initiating these pathobiologic responses to identify potential targets for therapeutic intervention. The premise underlying this approach is that, although the exact triggering mechanism(s) for acute pancreatitis caused by alcohol and gallstones in humans is not established, key steps in mediating the pathobiologic processes that define acute pancreatitis can be identified from animal models, and used to develop therapies that can be ultimately tested in human pancreatitis. In this article, we will first describe our current state of knowledge about the mechanisms of pancreatitis, which are largely based on studies in animal models, followed by a discussion of the clinical disorder and its management, especially related to issues that give rise to severe disease. Finally, in the last section, we will provide our perspective about future research directions needed to improve outcome in this group of patients. Abbreviations used in this paper: ERCP, endoscopic retrograde cholangiography; FAEE, fatty acid ethanol esters; FNA, fine needle aspiration; HSP, heat shock proteins; IAP, inhibitors of apoptosis; PAF, platelet-activating factor; PAR-2, proteinase-activated receptor-2; PARP, polyadp-ribose polymerase by the AGA Institute /07/$32.00 doi: /j.gastro

3 1056.e2 PANDOL ET AL GASTROENTEROLOGY Vol. 133, No. 3 Figure 1. Intracellular and extracellular factors and their influence on the pathobiologic processes of acute pancreatitis. This graphic presents the relationships between extracellular and intracellular factors underlying the mechanism of acute pancreatitis and the pathobiological processes of acute pancreatitis. Pathobiologic Processes Figure 1 represents an overview of processes that need to be considered. This figure illustrates unique extrapancreatic and intrapancreatic cellular events/factors involved in disease pathogenesis. Examples of extracellular factors include vascular and neural participation in pancreatitis. Examples of intracellular events/factors are activation of digestive enzymes, inhibition of secretion, cell calcium, heat shock proteins, inflammatory signaling pathways, and cell death pathways. Figure 1 suggests that these factors/events all have important roles in regulating inflammation, edema, and cell death responses as well as the systemic inflammatory response and multisystem organ failure in severe disease. Experimental Animal Models of Acute Pancreatitis There are 2 principal functions for animal model research in acute pancreatitis. These are investigations of the molecular mechanisms underlying the pathobiologic responses and testing of potential therapies before human trials. At present, only animal models provide the ability to reveal the sequence of initiating molecular steps resulting in the pathobiologic processes of acute pancreatitis. Moreover, there are considerable difficulties in designing human clinical trials related to the fact that the disease varies widely in course and severity. In addition, there is a low incidence of the most severe forms of pancreatitis in which testing agents for therapeutic benefit would have the most value. Thus, testing a large number of agents is not feasible. Animal models of acute pancreatitis can be used to screen potential therapies so that only the most promising ones advance to human testing. There have been several animal models of acute pancreatitis developed.2 9 The more commonly used models are listed in Table 1. The specific animal model selected for an experiment depends upon the specific goals of the experiment. For example, for testing a potential therapy, animal models with more severe pancreatitis and systemic inflammatory response should be used because this is the type of disease in which therapy will likely have the greatest benefit in humans. For studies designed to reveal the molecular signals underlying a particular pathobiologic response such as inflammation, cell death, or systemic inflammatory response, the most important issue in selection pertains to whether the particular animal model expresses the pathobiologic response under investigation. As Table 1 indicates, there are significant differences in the pathobiologic responses among species and among models in 1 species that must be considered to choose a model appropriate to address the response under investigation.8 10 A recent mouse model of acute

4 September 2007 PANCREATITIS 1056.e3 Table 1. Models of Experimental Acute Pancreatitis Models Species Features Cholecystokinin analogues (parenteral) Rat Pancreatic inflammation, apoptosis, mild necrosis. Systemic inflammation Cholecystokinin analogues (parenteral) Mouse Pancreatic inflammation, severe necrosis. Systemic inflammation Pancreatic duct obstruction Rat Pancreatic apoptosis Pancreatic duct obstruction Opossum Pancreatic inflammation, severe necrosis. Systemic inflammation Bile acid perfusion of pancreatic duct Rat Pancreatic inflammation, severe necrosis. Systemic inflammation Choline-deficient, ethionine-supplemented diet Mouse Pancreatic inflammation, severe necrosis. Systemic inflammation Arginine (parenteral) Rat Pancreatic inflammation, severe necrosis. Systemic inflammation Alcohol diet and cholesystokinin analogues (parenteral) Rat Pancreatic inflammation, mild necrosis. Systemic inflammation necrotizing pancreatitis has been introduced by the group of Saluja using intraperitoneal applications of L-arginine. 11 Of particular note, none of the models are caused by factors that we currently believe cause human pancreatitis. Thus, considerable caution should be used when applying the results from experimental animal models to designing therapies for humans. A judicious approach should include focus on treating a specific pathobiologic response; showing that the mechanism of the pathobiologic response is similar across several animal models; and that a proposed treatment is effective in attenuating the pathobiological response across several animal models. Of note, because genetically modified mice are becoming widely available, experimental models of acute pancreatitis are now often used to test the participation of specific biochemical pathways. 4 Pathobiologic Inflammation and Edema Responses It is generally believed that the severity of pancreatitis might be determined by events that occur after acinar cell injury. Pancreatic acinar cells also synthesize and release cytokines and chemokines, resulting in the recruitment of inflammatory cells such as neutrophils and macrophages. Recruitment and activation of various inflammatory cells leads to further acinar cell injury and causes an elevation of various proinflammatory mediators such as tumor necrosis factor- (TNF- ), interleukin (IL)-1, IL-2, IL-6, and other chemokines and anti-inflammatory factors such as IL-10 and IL-1 receptor antagonist. 12,13 The role of some of these cytokines in disease severity has been demonstrated for some including TNF- and IL-1 using animals with genetic deletions for receptors for these cytokines. 14 These inflammatory cells and mediators play a role in the systemic manifestations besides modulating pancreatic acinar cell injury. In most patients, the acinar cell damage and local inflammation associated with pancreatitis will resolve, but, in some cases, the disease progresses to systemic illness. Systemic inflammatory response syndrome (SIRS) is a result of uncontrolled local inflammation and predisposes to multiple organ failure. Pancreatitis-associated lung injury (adult respiratory distress syndrome [ARDS]) is frequently a factor in the early death of patients with severe acute pancreatitis. 15 Several key inflammatory mediators including plateletactivating factor (PAF) and substance P have been identified to play a key role in experimental cholecystokinin (CCK)-induced pancreatitis as well as in human disease. Inhibition of PAF by either accelerating its degradation by recombinant PAF acetylhydrolase 16 or PAF antagonists results in amelioration of the severity of pancreatitis. The neuropeptide substance P, which is released from afferent nerve endings and is important in inflammatory processes, 21 has also been shown to be of significance in regulating the severity of pancreatitis. Caerulein-induced pancreatitis is reduced in severity in mice lacking substance P receptors (neurokinin 1 receptor), 22 as well as by pretreatment with the neurokinin 1 receptor antagonist (CP-96345). 23 The clinical trials evaluating the efficacy of the PAF receptor antagonist (lexipafant) have produced conflicting results, suggesting that more than one mediator might be at work and that a combination therapy blocking more than a single mediator might be needed to produce substantial results A hallmark of acute pancreatitis is the accumulation of neutrophils in the pancreas, which have been shown to modulate the severity of both the local changes as well as the systemic manifestation of pancreatitis. Depletion of neutrophils with antineutrophil serum leads to attenuation of severity of pancreatitis and associated lung injury. 27,28 In addition to the involvement of extracellular signaling in the inflammatory response described above, mechanisms of intracellular signaling have also been identified that mediate the inflammatory response in the pancreas and are system-

5 1056.e4 PANDOL ET AL GASTROENTEROLOGY Vol. 133, No. 3 ically These signaling systems include nuclear factor- B (NF- B) activator protein-1 (AP-1), MAPK-modules, Stat3, and phosphatidylinositol-3 kinase (PI-3 kinase). Inhibition of these signals has been demonstrated to decrease inflammation and improve the severity of pancreatitis in most cases ,34 Studies of NF- B have demonstrated that it is activated in the pancreatic acinar cell early before the influx of inflammatory cells into the tissue. 32 Intraductal adenoviral gene transfer of the activating subunit RelAp65 was sufficient to induce acute pancreatitis with systemic complications. 34 However, pharmacological strategies to inhibit NF- B activation revealed mixed results. Most studies suggest a harmful role of NF- B activation with one exception proposing a protective role. 35 Furthermore, pharmalogical inhibition of MAPK-modules during the onset of acute pancreatitis resulted in mixed outcomes as well. 36,37 Interestingly, blocking PI3-Kinase activation resulted in amelioration of pancreatitis in 2 rodent models without affecting NF- B activation. The interaction and cross talk of these signaling pathways is not well understood. Thus, a better understanding of the signaling network is mandatory for the identification of therapeutic targets. Activation of acinar cell CCK and TNF- receptors through effects on protein kinase C isoforms can cause the NF- B activation in acinar cells In addition to up-regulating proinflammatory cytokines, NF- B can increase adhesion molecule ICAM-1 effect in the pancreas. 43 NF- B is also involved in the systemic inflammatory response of acute pancreatitis. The involvement of the pancreatic enzymes themselves in the systemic inflammatory response of acute pancreatitis has been proposed. For example, pancreatic elastase but not amylase and lipase causes the pulmonary injury of the adult respiratory distress syndrome by activating NF- B in pulmonary tissue. 44 Pancreatic elastase causes liver injury through increasing cytokine production in Kupffer cells utilizing NF- B signaling. 45 However, recent data question this concept since various elastase preparations were found to be contaminated with LPS, and low endotoxin elastase failed to induce proinflammatory effects in vivo and in vitro. 46 Although less well investigated, stress kinases Erk and p38 are also activated in experimental pancreatitis and may play a role in pathogenesis Of particular interest are observations that during pancreatitis, there is up-regulation of an anti-inflammatory system that is probably protective. This system includes a transcriptional regulator called p8, and it is one of its regulated genes, pancreatitis-associated protein I (PAPI). 50,51 Both p8 and PAPI are rapidly induced during pancreatitis. Animals with genetic deletions of p8 have a markedly augmented pancreatic inflammatory response during pancreatitis, 50 and antibodies to PAPI injected into animals with experimental pancreatitis also markedly augmented pancreatic inflammation. 50 These findings suggest that PAPI is an anti-inflammatory cytokine similar to IL-10. Pathobiologic Cell Death Responses: Apoptosis and Necrosis Mechanisms of cell death in acute pancreatitis are emerging as critical topics for investigation. There are 2 reasons for this. First and most importantly, pancreatic necrosis during acute pancreatitis is a key factor predictive of outcome. 8 10,29 33 The second reason is that the fundamental biology underlying cell death responses has been rapidly advancing in recent years so that new concepts and techniques are being developed that can be applied to the study of pancreatitis. There is general consensus that strategies aimed at inhibiting or preventing necrosis during pancreatitis would improve outcome in patients with severe pancreatitis. 8 10,52 54 Thus, identification of methods to inhibit necrosis or switch necrosis to apoptosis during acute pancreatitis will likely improve outcome. The following subsection provides a general description of cell death mechanisms as well as findings specific to the pancreas during acute pancreatitis. There are several cellular components involved in cell death responses. Figure 2 is used to illustrate mechanisms relevant to the pancreas and the pancreatic acinar cell. There are 2 major classifications of cell death that have been investigated in experimental acute pancreatitis models: apoptosis and necrosis. Apoptosis is also called programmed cell death and is required for normal development and tissue homeostasis but can also occur in pathologic conditions An apoptotic cell undergoes a cascade of molecular events ultimately leading to its total removal. These events result in morphologic changes that include shrinkage of the cell and its organelles and condensation of nuclear chromatin. The nuclear chromatin changes are accompanied by intranucleosomal DNA cleavage that can be identified on DNA gels as a laddering effect. The molecular signals of apoptosis also lead to translocation of phosphatidylserine from the inner leaflet of the plasma membrane to the outer leaflet. The shrunken apoptotic cell often referred to as an apoptotic body with its externally located phosphatidylserine is recognized by macrophages that phagocytize the apoptotic body and remove it from the tissue. With apoptosis, there is no leakage of cellular contents to the extracellular environment and minimal inflammation. Necrosis exclusively occurs in pathologic situations and represents a severe cellular response to injury The events of necrosis include mitochondrial swelling and dysfunction, rupture of the plasma membrane, and release of intracellular constituents to the extracellular space. Necrosis is associated with an acute inflammatory response. The DNA degradation is less extensive with necrosis than with apoptosis, and there is no laddering effect on DNA gels because the cleavage of DNA is irregular. Information about apoptosis and necrosis during acute pancreatitis is now emerging. 8 10,38,58,61 64 In gen-

6 September 2007 PANCREATITIS 1056.e5 Figure 2. Intracellular mechanisms involved in the death response of pancreatic tissue during acute pancreatitis. The slide depicts both the interrelationship between apoptosis and necrosis pathways and various regulatory systems that may alter the outcome of cell death in pancreatitis. The illustration indicates that in addition to caspase and mitochondrial pathways mediating cell death, endosplasmic reticulum stress pathways can mediate cell death at least in part through release of its stored calcium that can activate mitochondrial death pathways. Lysomal pathways are involved through their cathepsins, which can mediate cell death in many tissues. In the pancreas, cathepsins can also activate digestive enzymes in the cell, which can augment cell damage. Severe mitochondrial dysfunction can lead to ATP depletion, and this de-energinization of the cell can lead to necrosis. In this case, apoptosis will not occur because ATP is necessary for the activation of effector caspases. As shown in the slide, the enzyme, polyadp-ribose polymerase (PARP) decreases ATP. PARP is up-regulated in cell injury where it is involved in DNA repair. PARP utilizes ATP in the repair process resulting in depletion of ATP and further cell vulnerability to necrosis. Phosphatidylinositol 3-kinase (PI 3-kinase), nuclear factor- B (NF- B) and inhibitors of apoptosis (IAPs) can inhibit apoptosis, a mechanism of importance in cancer as well as pancreatitis. This inhibitory system represents one pathway for cancer cell resistance to death. Finally, effector caspases can have several effects as indicated in the illustration to modulate the cell injury and death response. eral, apoptosis in pancreatic tissue (acinar cells) like nonpancreatic tissues is mediated mainly by activation of a family of cysteine proteases, namely, caspases. Caspases are synthesized as inactive precursors and are activated by cleavage induced by proteases including the caspases themselves. Caspases are divided into 2 major classifications. These are effector or executioner caspases, which are responsible for most of the cleavages that disassemble the cell, and initiator caspases, which trigger cleavage and activation of the executioner caspases. There is mitochondrial involvement in apoptosis as well. Stresses to the cell can act directly on the mitochondria or though initiator caspases to cause specific mitochondrial changes that include the following: depolarization of the mitochondrial membrane potential; permeabilization of the outer mitochondrial membrane; and release of factors into the cytoplasm that interact with the caspase system (apoptogenic factors). A major mitochondrial apoptogenic factor is cytochrome C. The mitochondrial outer membrane permeabilization is thought to occur through opening of the mitochondrial membrane permeability transition pore (PTP). PTP inhibitors such as cyclosporin A are able to prevent cytochrome C release, thus inhibiting caspase activation and apoptosis in many cells including pancreatic acinar cells from rats. 61,65 Cholecystokinin-octapeptide at supraphysiologic doses, tumor necrosis factor-, and lipopolysaccharide have been shown to cause apoptosis of pancreatic acinar cells in vitro. 10,58,61,63,64 Activation of initiator and effector caspases as well as opening of the mitochondrial PTP with resulting cytochrome C release has been determined to be necessary for pancreatic acinar cell apoptosis. 10,62 The members of the Bcl-2 family regulate mitochondrial permeability. 56,57,66 68 Some, such as Bcl-2 and BclxL, inhibit mitochondrial permeability; whereas others, such as Bax, Bak, Bad, Bid, and Bcl-xS, enhance permeability. In pancreatitis there are alterations in the expression of some of these proteins. However, the significance of these alterations is unknown. 69,70 The mechanisms underlying necrosis are beginning to be explored. There is a general consensus that the level of adenosine triphosphate (ATP) in the cell can determine

7 1056.e6 PANDOL ET AL GASTROENTEROLOGY Vol. 133, No. 3 whether a cell will die by apoptosis or necrosis. 58,71 74 For example, greater levels of ATP depletion with a pancreatic stress stimulus will likely result in necrosis, whereas lesser amounts of ATP depletion will result in apoptosis. Because mitochondria are the principal generators of ATP production, necrosis results when mitochondrial injury is greater, resulting in more profound changes in mitochondrial function than those occurring with apoptosis. One reason for the effect of low ATP leading to necrosis is that ATP is necessary for key steps in the caspase pathway. If the caspase pathways and apoptosis are blocked, the cell will likely proceed to necrosis. 10 Another importance of ATP is that it is necessary for maintaining ion pumps that are essential for cell integrity. 58,71 74 Thus, the level of mitochondrial function during acute pancreatitis is one key arbitrator of the cell death type. There are potentially several other regulators of cell death type with pancreatic stress as depicted in Figure 2. These include polyadp-ribose polymerase (PARP), inhibitors of apoptosis (IAPs), PI-3 kinase, NF- B, calcium, cathepsin B, and trypsin. PARP is activated by DNA breaks during cell injury, and its activation leads to ATP depletion and necrosis because ATP is utilized during the process. 75,76 Studies in experimental models demonstrate that PARP activity markedly increases during acute pancreatitis and that pharmacologic blockade of PARP decreases necrosis A recent study demonstrated the importance of IAPs in regulating the type of death response by determining the mechanisms accounting for differences in cell death responses between experimental models of pancreatitis. 10 In the pancreas, the IAP called XIAP was rapidly degraded in models of experimental acute pancreatitis in which caspase activation and apoptosis predominate; however, XIAP was not degraded and caspases were inhibited in models in which necrosis predominates. Furthermore, interventions that blocked XIAP resulted in increased caspase activation and necrosis. The conversion of the cell death response to apoptosis was associated with improvement in the severity of pancreatitis. Signals involved in the inflammatory response such as NF- B and PI-3 kinase have been shown to be involved in the regulation of the expression of IAPs in other tissues Furthermore, inhibition of both NF- B and PI-3 kinase activation has been demonstrated to decrease pancreatitis severity including decreased necrosis and increased apoptosis in experimental animal models of acute pancreatitis ,33,34 Thus, although not tested, NF- B and PI-3 kinase systems could be involved in the promotion of necrosis through their effects to increase expression of IAPs. Sustained high levels of cytoplasmic free calcium concentrations in the acinar cell are another regulator of cell death during pancreatitis. In the resting cell, calcium is largely sequestered in the endoplasmic reticulum, and the cytoplasmic free calcium concentrations are maintained at approximately 100 nmol/l. 86 Hormones and neurotransmitters release the calcium from the endoplasmic reticulum into the cytoplasm of the cell by their ability to increase inositol 1,4,5-trisphosphate. 87 Under physiologic conditions, the changes in calcium concentrations are very transient, and the transient changes mediate normal responses such as secretion of proteins and ions. However, very high concentrations of CCK analogues or cholinergic agonists cause larger and more sustained increases in the cytoplasmic calcium concentrations. The excess cytoplasmic calcium in this situation is taken up by the mitochondria, which leads to mitochondrial dysfunction and cell death. The greater the concentrations of cytosolic calcium and the longer the duration of the increase, the more likely that the cell death will be from necrosis Of note, bile acids and fatty acid ethyl esters, potential initiators of gallstone and alcohol pancreatitis, respectively, cause cytoplasmic calcium increases leading to necrosis through mitochondrial dysfunction, suggesting a key role for calcium in cell death responses. The cysteine protease cathepsin B is a major lysosomal enzyme in the pancreas. Cathepsins have been demonstrated to play a role in both apoptosis and necrosis in a variety of cell types. 58,60 Specific inhibition of cathepsin B significantly decreases necrosis but not apoptosis in experimental models of pancreatitis. 96,97 As described above, part of the effect of cathepsin B may be mediated by its effect to catalyze the conversion of trypsinogen to its active form. An interesting observation is that the caspases mediate cleavage of PARP, which results in decreased PARP activity and inhibits trypsin activity. 58,62 Such results suggest that, once apoptosis is initiated, the cell provides mechanisms to inhibit trypsin activity and necrosis so that the cell can proceed through apoptotic cell death without interruption. One additional arbitrator of the death response in experimental models of acute pancreatitis is the inflammatory response itself ,65,98 For example, removal of neutrophils from animals during experimental pancreatitis markedly decreases pancreatic necrosis and the severity of pancreatitis while increasing pancreatic apoptosis. The effect of neutrophils is likely mediated by both their production of reactive oxygen species and secretion of elastase. 27,99 As indicated earlier, inflammatory signals such as NF- B and PI-3 kinase likely affect death signaling within the acinar cell. Thus, both inflammatory signaling in the pancreatic parenchymal cells and the infiltration of inflammatory cells have affects to promote necrosis. Intracellular Factors: Digestive Enzyme Activation, Inhibition of Secretion and Calcium The acinar cells elaborate a plethora of digestive enzymes that are synthesized and stored as inactive zy-

8 September 2007 PANCREATITIS 1056.e7 mogen precursors to avoid autodigestion. Moreover, intracellular protective mechanisms such as the presence of trypsin inhibitor, nonoptimal ph conditions, and proteases degrading the activated enzymes normally can prevent cell damage by low levels of activated zymogen. Various in vitro and in vivo studies have established beyond a doubt that the intra-acinar activation of zymogen is a key event in the pathogenesis of pancreatitis. Activation of trypsinogen and other zymogens has been observed in the pancreatic homogenate as early as 10 minutes after supramaximal stimulation by caerulein in rats and increases over time The fact that other markers of pancreatitis, eg, hyperamylasemia, pancreatic edema, and acinar cell vacuolization, can only be detected 30 minutes after supramaximal stimulation 101 strongly supports the paradigm that zymogen activation is the cause and not the result of pancreatitis. Similarly in in vitro studies, supramaximal stimulation by CCK ( mol/l) results in zymogen activation and acinar cell injury. 104,105 Moreover, pretreatment with protease inhibitors has been shown to reduce the severity of hyperstimulation pancreatitis in animals and prevents acinar cell injury These findings emphasize the role of zymogen activation in acinar cell injury caused by supramaximal secretagogue stimulation. It appears that, for the pathogenesis of pancreatitis, not only do zymogens need to be activated but they must also be retained inside the acinar cells. A comparison of the dose-response curves of the various secretagogues provides evidence for this conclusion. It has been observed that the dose-response curve of caerulein for pancreatic secretion is biphasic, with maximal stimulation at 0.2 g/kg per hour 107 ; increasing the dose to 5 g/kg per hour paradoxically inhibits the secretion. Similar doseresponse curves were obtained for caerulein and carbacholamine in vitro. 107,108 Only this supramaximal dose of caerulein induces pancreatitis as defined by hyperamylasemia, pancreatic edema, and acinar cell injury in vivo. 107,108 Experiments with CCK-JMV-180 (an analogue of CCK) and bombesin further support the role of inhibition of secretion and retention of secretion in the pathogenesis of pancreatitis. CCK-JMV-180, in contrast to caerulein, has a monophasic dose-response curve for amylase secretion and does not inhibit secretion at high doses. 107 Remarkably, a supramaximal dose of JMV-180 does not cause pancreatitis. 107 Similarly, bombesin also has a monophasic dose-response curve but, unlike that of caerulein, does not induce pancreatitis at supramaximal doses (500 g/kg per hour). 104,109 The role of inhibition of secretion and retention of the activated zymogens is further supported by studying the effects of proteinase-activated receptor-2 (PAR-2) activation in pancreatitis. Pancreatic acinar cells express the G-protein-coupled transmembrane receptor known as PAR-2, which can be activated by trypsin. PAR-2 has been shown to be activated (probably by trypsin, which is prematurely activated and released into the interstitial space) during acute pancreatitis. 110 Administration of the PAR-2-activating peptide SLIGRL decreases the severity of CCK-induced pancreatitis. 111,112 Furthermore, injury from CCK-induced pancreatitis is more extensive in PAR-2 knockout mice compared with the wild-type animals, 113 suggesting that PAR-2 activation is protective in pancreatitis. The protective effect of PAR-2 activation in pancreatitis is believed to be due to reversal of the block in secretion observed in both the caerulein- and arginineinduced pancreatitis mouse models. 29 Several factors are involved in zymogen activation and inhibition of secretion that are central to the pathogenesis of pancreatitis. The intra-acinar actin web is involved in the secretory function of these cells. Supramaximal doses of caerulein induce a loss of the terminal actin web and its associated intermediate filaments, 114,115 which are believed to be responsible for the inhibition of secretion observed at supramaximal doses. CCK-JMV-180, on the other hand, causes neither ultrastructural change in the acinar subapical actin web, and as discussed above, nor does it lead to inhibition of secretion at supramaximal concentration. 116 The mechanism by which supramaximal doses of caerulein leads to ablation of the actin cytoskeleton is unclear, although a role has been suggested for the activation of phospholipase C by stimulation of low-affinity receptors. 116 Inhibition of secretion may also be due to disorders of the exocytotic process related to SNARE proteins and small GTP binding proteins. 117,118 Specific SNARE proteins located on the plasma membrane and the zymogen granule membrane regulate exocytosis through their interactions. Investigations in pancreatic acinar cell demonstrate high-dose CCK-8 causes displacement of one of the SNARE proteins, Munc 18c, from the basal surface of the acinar cell with a concomitant redirection of apical exocytosis to the basal surface. 117 This displacement has been observed in human alcoholic pancreatitis. 119 Small GTP binding proteins of the Rab family have been demonstrated to also have roles in the exocytosis process These proteins are a large family involved in vesicular traffic and membrane fusion in eukaryotic cells. In the exocrine pancreas, Rab 3b and Rab 27b are present as zymogen granule membranes and play a necessary role in exocytosis Many hypotheses have been proposed for the mechanism of intra-acinar activation of zymogens. One intriguing theory is the colocalization hypothesis, which suggests that, during the early stages of pancreatitis, pancreatic zymogens become colocalized with lysosomal enzymes and thus become activated by lysosomal enzymes such as cathepsin B to produce active trypsin. This in turn activates other digestive zymogens, resulting in cell injury. The colocalization of lysosomes and zymogen granules has been observed using subcellular fractionation 123 and immunolocalization studies. 124 This colocalization was noted within 15 minutes of

9 1056.e8 PANDOL ET AL GASTROENTEROLOGY Vol. 133, No. 3 the start of caerulein administration, and trypsinogen activation was observed at the same time. 101 By contrast, hyperamylasemia, pancreatic edema, and acinar cell injury were observed at later time points, which suggests that colocalization is not the result of cell injury, but the cause of it. Colocalization is not a model-specific phenomenon but has been observed in other models of pancreatitis and might be applicable to human disease. However, some concerns have been expressed about the colocalization hypothesis with regard to the ability of cathepsin B to activate trypsinogen because the optimal ph of the lysosomal enzymes is approximately 5, whereas the ph at the sites of colocalization appeared to be neutral. However, studies have now demonstrated that cathepsin B has substantial ability to activate trypsinogen even at neutral ph In vivo as well as in vitro experiments have further supported the role of lysosomal enzymes in the activation of trypsinogen. Pretreatment of rat pancreatic acini with E-64d, which is a cell-permeable form of the cathepsin B inhibitor E-64, completely inhibited cathepsin B and prevented caerulein-induced activation of trypsinogen. 131 Notably, the concentrations of E-64d that led to partial inhibition of cathepsin B in vitro could not prevent caerulein-induced trypsinogen activation. Similarly, treatment of the acini with another cathepsin B inhibitor, CA-074me, completely prevented caerulein-induced in vitro activation of trypsinogen activation. 97 In animals, inhibition of cathepsin B by CA-074me reduced the severity of caerulein-induced pancreatitis. The role of cathepsin B in the activation of trypsinogen has been put to test by the use of cathepsin B knockout mice as well. After induction of experimental secretagogue-induced pancreatitis, the trypsin activity in the pancreas of cathepsin B knockout mice was more than 80% lower than in the wild-type animals. 96 Also, pancreatic damage as indicated by various parameters including the extent of acinar tissue necrosis was substantially lower in the knockout animals. The above-mentioned studies have definitively proven the roles of colocalization and lysosomal enzymes in intrapancreatic trypsinogen activation and the onset of acute pancreatitis. Activation of PI-3 kinase (class 3) has been proposed as the mechanism of colocalization because inhibition of PI-3 kinase activity by either wortmannin or LY prevents the colocalization of zymogen and lysosomal hydrolases and thus decreases activation of trypsinogen and decreases the severity of pancreatitis in caerulein pancreatitis as well as in a model of pancreatitis involving injection of taurocholate into the bile duct. 29 Additional mechanisms have been proposed for the activation of trypsinogen. For example, neutrophils, which generally are believed to modulate the late events in pancreatitis, are said to have some role in trypsinogen activation. 27,99 Also, trypsin activiation is reduced 27,99 by inhibition of neutrophil elastase, which is needed to cleave cell contacts between acinar cells and is thus important for neutrophil infiltration or depletion of the neutrophils by antineutrophil serum. Substantial direct and indirect evidence exists for the involvement of calcium in zymogen activation and the pathogenesis of pancreatitis as mentioned previously. In animal experiments, hypercalcemia can either decrease the threshold level for the onset of pancreatitis or can induce morphologic alterations resembling pancreatitis. Disturbances in the calcium homeostasis of pancreatic acinar cells have been observed early in the secretagogueinduced model of pancreatitis. 132 Moreover, attenuation of calcium elevation in acinar cells by the cytosolic calcium chelator BAPTA-AM (1,2-bis-[o-Aminophenoxy]- ethane-n,n,n=,n=-tetraacetic acid, tetracetoxymethyl ester) prevents zymogen activation, again suggesting that calcium is essential for zymogen activation. 105,133,134 It has been shown that it is not the initial transient rise in calcium but the sustained elevation following the initial spike induced by caerulein that is responsible for zymogen activation. 104,105 In the absence of extracellular calcium, both the sustained plateau phase of calcium as well as the activation of trypsinogen induced by a supramaximal dose of caerulein is attenuated. Thus, it is evident that calcium is essential for zymogen activation as well as pancreatitis, but the controversial issue is whether calcium in itself is sufficient to induce zymogen activation and pancreatitis. Some investigators believe that elevation of intra-acinar calcium regardless of its source or its stimulatory agent (eg, thapsigargin, ionomycin) is sufficient by itself to induce protease activation. 133 However, studies involving elevation of intra-acinar calcium by a variety of reagents with different sites of action have shown that mere elevation of intracellular calcium does not lead to trypsinogen activation. This suggests that calcium, although essential, is not sufficient by itself to activate trypsinogen. Presumably, then, other effects of supramaximal secretagogue are needed to activate trypsinogen and induce pancreatitis. Intracellular Factors: Heat Shock Proteins The family of proteins that provide cellular protection against the toxic mediators of inflammation are called heat shock proteins (HSP). 135,136 Synthesis of these proteins when a cell is under stress is mediated by the transcriptional activation of heat shock factor-1(hsf- 1). 137 HSPs in the pancreas are induced by stresses such as heat shock, water immersion, hyperosmolarity, chemicals, or supramaximal doses of CCK. HSPs are up-regulated in the pancreas after administration of supramaximal doses of caerulein and protect against injury. The synthesis of both HSP27 and HSP70 is increased in isolated pancreatic acini as well as in the pancreas of animals with caerulein-induced pancreatitis HSPs are increased in other models of pancreatitis as well, including the dibutyltin dichloride

10 September 2007 PANCREATITIS 1056.e9 Figure 3. Etiologies of acute pancreatitis. The pie graph represents the relative frequencies of the etiologies of acute pancreatitis. Other features of the pancreatitis resulting from these etiologies are discussed in the text. model 141 and the arginine-induced model. 142 Evidence that HSPs are protective is provided by the fact that caerulein-induced pancreatitis is more severe in HSF-1 knockout mice that are incapable of HSP synthesis 143 ; over expression of HSP27 protects against cerulein-induced pancreatitis. 144,144a Further support that induction of HSPs is a self-defense mechanism comes from studies examining the effect of HSP over expression on the severity of pancreatitis. Prior induction of HSP by water immersion (HSP60 and HSP70) or hyperthermia (HSP70) before caerulein administration leads to decreased severity of pancreatitis. 141, Over expression of HSPs by other means including ischemic preconditioning, 150 the HSP70 coinducer BRX-220, 151 or transgenic over expression 144,144a has also been shown to protect against pancreatitis. The first direct evidence that HSP70 plays an essential role in thermal stress-induced protection against pancreatitis was provided by Bhagat et al, 136 who used antisense against HSP70. Antisense but not sense-hsp70 reduced thermal stress-induced HSP70 expression, restored the ability of supramaximal caerulein stimulation to cause intrapancreatic trypsinogen activation, and abolished the protective effect of prior thermal stress against pancreatitis. 138 A clue to how HSP70 protects in pancreatitis was provided in a recent article by Hwang et al, 152 showing that basal trypsin activity and the zymogen/lysosomal ratio of cathepsin B activity before caerulein injection was higher in HSP70 knockout mice compared with wild-type animals. Therefore, lysosomal enzyme/digestive zymogen colocalization is prevented by HSP70, intra-acinar cell trypsinogen activation is reduced, and, consequently, pancreatitis is ameliorated. HSP70 may reduce pancreatic injury by blocking the intracellular trafficking changes that lead to zymogen activation and pancreatitis and preventing the pathologic rise in calcium required for trypsinogen activation. In addition, HSPs could also affect other inflammatory mediators such as NF- B. 144,144a,153 Extracellular Factors: Neuroinflammatory and Neurovascular Responses Neural involvement in experimental acute pancreatitis models has best been characterized for an intrapancreatic neural system that involves activation of the transient receptor potential vanilloid-1 receptor on sensory neurons in the pancreas. 22, The activation of these receptors stimulates release of substance P and calcitonin gene-related peptide (CGRP) from neural endings. Substance P and CGRP, in turn, mediate both increased vascular permeability resulting in edema and neutrophil infiltration into the parenchyma of the pancreas. Blockade of either transient receptor potential vanilloid-1 receptor activation or neurokinin-1 receptors that mediate the effects of substance P significantly improves the severity of experimental pancreatitis. There are vascular disorders of both larger vessels and microcirculation of the pancreas during acute pancreatitis. The role of vascular abnormalities in mediating the severity of acute pancreatitis in humans was underscored by a report, 54 showing significant vascular changes occurring early in the course of cases that went on to have severe necrotizing pancreatitis. There was a high incidence of vasospasm (both intrapancreatic and extrapancreatic) detected by angiography in those patients who went on to have necrotizing pancreatitis as measured by contrast-enhanced CT studies. In fact, areas of necrosis often corresponded to sites of vasospasm. The microvas- Figure 4. Pathways of ethanol metabolism in the pancreas. The majority of ethanol is metabolized by the liver by the oxidative system in the liver. A small amount of oxidative ethanol metabolism takes place in the pancreas. On the other hand, nonoxidative ethanol metabolism is prominent in the pancreas resulting in the formation of fatty acid ethanol esters. These esters have been demonstrated to have several pathologic effects on the cell biology of pancreatic parenchymal cells.

11 1056.e10 PANDOL ET AL GASTROENTEROLOGY Vol. 133, No. 3 cular circulation is also significantly affected in acute pancreatitis. Studies of microcirculation of the exocrine pancreas show that, in the normal state, the pancreatic acinus is surrounded by a large number of small capillaries of small diameter forming a network surrounding the acinus. 158,159 This capillary network is responsible for edema and transmigration of inflammatory cells during pancreatitis. Furthermore, decreases in microperfusion occur in more severe forms of experimental pancreatitis, suggesting that microperfusion is a key event in necrosis. In addition to substance P, there are several other mediators that are likely involved in the microperfusion of severe pancreatitis These include endothelin-1 and endothelial nitric oxide synthase as well as cytokines and chemokines involved in the inflammatory response. Causes of Pancreatitis Figure 3 shows the relative distribution of the conditions associated with acute pancreatitis. We will now discuss our understanding of several of the most important conditions. Of note, autoimmune pancreatitis will be discussed in a future review on chronic pancreatitis. Alcohol Abuse and Acute Pancreatitis Alcohol abuse is a major cause of acute pancreatitis as well as the leading cause of chronic pancreatitis. Although the incidence of pancreatic disease increases as a function of the prevalence of alcohol abuse in a population, only a minority of subjects who abuse alcohol develop pancreatitis. 5,7 Also, for the most part, alcohol feeding does not lead to pancreatitis in animals. 5 These findings have led to a working hypothesis that alcohol abuse sensitizes individuals to pancreatitis caused by pancreatic stress. Examples of pancreatic sensitization to pancreatitis as well as the pathobiologic responses inflammation, necrosis, and trypsin activation have been revealed in experimental animal models. 42, There are several effects of ethanol that likely underlie these effects of alcohol on the pathobiologic responses. First, ethanol has effects on the inflammatory signaling systems in the pancreatic acinar cell. 42,175,178 For example, ethanol increases the sensitivity of the pancreatic acinar cell to activation of the inflammatory signal NF- B through effects on kinases that regulate NF- B activation. These kinase effects have been narrowed to specific isoforms of protein kinase C. 142 Ethanol feeding in rats results in a decrease in the expression and/or activity of both initiator and executioner caspases, likely because of inhibition of the Janus kinase 2/signal transducers and activators of transcription 1 (JAK2/STAT 1) signaling pathway in the pancreatic parenchyma. 179,180 By contrast, ethanol feeding enhances the expression and activity of cathepsin B. 179 The importance of cathepsin B is that it may enhance necrosis in experimental pancreatitis, possibly through its ability to catalyze conversion of trypsinogen to active trypsin in the pancreatic acinar cell. 95,96 The combination of these effects is likely to result in increased necrosis induced with pancreatic stress, such as duct ligation 9 or lipopolysaccharide. 180 Ethanol in the absence of pancreatitis causes decreased microperfusion of the exocrine pancreas, and ethanol augments the decreased microperfusion caused by endothelin ,181 These effects of ethanol likely represent another mechanism of ethanol abuse that sensitizes the pancreas to necrotizing pancreatitis. 182 The aforementioned effects of alcohol on the pancreas do not consider the unique metabolism of ethanol in the pancreas (Figure 4). In the liver, the major ethanol-metabolizing organ in the body, most ethanol is metabolized by an oxidative system converting ethanol to acetaldehyde, and then to acetate using the enzymes alcohol dehydrogenase and acetaldehyde dehydrogenase. 183,184 An alternative pathway for ethanol metabolism is through the transient formation of fatty acid ethanol esters (FAEEs) that occurs in several organs including the pancreas Compared with the liver, the activity of FAEE synthesis is greater in the pancreas. 186, In contrast, oxidative metabolism of ethanol in the pancreas is much less than that in the liver. FAEEs have been found in high concentrations (ie, up to 150 mol/l) in the pancreas in postmortem examinations of patients dying while intoxicated. 186 Furthermore, FAEEs have been measured in human serum up to concentrations of 41 mol/l. 186,187 FAEEs are produced in the pancreatic acinar cell by fatty acid ethyl ester synthases. 175, The reason for considerable interest in FAEEs in recent years with respect to pancreatitis is that they can produce some of the pathobiologic effects of pancreatitis in animal models, including edema and intracellular typsin activation. 138, Most interestingly, recent studies demonstrate that FAEEs cause sustained increases in the cytosolic calcium concentration in pancreatic acinar cells by activating inositol trisphosphate receptors in the cell, leading to mitochondrial injury (ie, through sustained cytosolic calcium increases) and impaired ATP production. 92,133,134, This sequence of events leads to necrosis of the pancreatic acinar cell. In summary, there are multiple effects of ethanol and its metabolites on the pancreas that are involved in sensitizing the pancreas to pancreatitis. These include effects on inflammatory and cell death signaling that can promote inflammation and necrosis. These findings from animal models are consistent with a recent report indicating that alcohol abuse in humans is a risk factor for pancreatic necrosis during pancreatitis. 182 Biliary Pancreatitis Over a century ago, Opie proposed that gallstones cause pancreatitis by impacting in the ampulla of Vater. 199 One result of this event would be reflux of bile into the pancreatic duct (Opie 1) as illustrated in Figure

12 September 2007 PANCREATITIS 1056.e11 different types of death responses. Of note, obstruction of the pancreatic duct in the rat leads to altered calcium dynamics in the acinar cell, indicating that calcium plays a role in the response to obstruction. 201 Figure 5. Mechanisms of gallstone pancreatitis. This graphic illustrates the 2 proposed mechanisms of initiation of gallstone pancreatitis. In Opie 1, the obstructing stone creates a common channel and increased pressure between the common bile duct and the pancreatic duct. This results in the potential for reflux of biliary contents into the pancreas. In Opie 2, only obstruction of the ducts occurs. Both mechanisms have the potential for leading to pancreatitis in humans. 5. Multiple studies in animals have shown that ductal perfusion with bile salts will cause pancreatitis in animals. 31 Furthermore, recent studies have demonstrated that the effects of bile acids are due to their uptake by an apically located bile acid transport, followed by both stimulation of calcium release from endoplasmic reticulum store of calcium through effects on inositol 1,4,5- trisphosphate (IP 3 ) receptors as well as prevention of re-uptake into the store by inhibition of the endoplasmic reticulum calcium uptake pump. 93,95 The resulting large increases in cytoplasmic calcium lead to mitochondrial uptake and dysfunction, followed by necrotic cell death. The changes in calcium may also cause inflammation by mediating activation of NF- B. 31,94 It should be noted that roles for bile acids in human pancreatitis have not been demonstrated yet. Pancreatitis from gallstones could also develop by obstruction of the pancreatic duct in the absence of bile reflux into the duct. Although there has been considerable debate about this possibility because duct obstruction in rats results in apoptosis of pancreatic cell and pancreatic atrophy with little or no pancreatitis. 9 Nevertheless, this effect may be due to differences in death response signaling among species. For example, obstruction of the pancreatic duct without bile reflux results in necrotizing pancreatitis in the opossum. 9,200 The differences in death signaling pathways between rat opossum and human pancreas are not known. Whether obstruction of the pancreatic duct without bile reflux causes necrotizing pancreatitis in humans is unknown. However, as previously pointed out, the apoptosis and necrosis death pathways are interconnected, and differences in signal responses to noxious stimuli certainly account for Drug-Induced Pancreatitis A variety of drugs (including but not limited to sulfonamides, tetracyclines, l-asparaginase, vinca alkaloids, diuretics, salicylates, and valproic acid) has been implicated in the etiology of pancreatitis. 201 There has been a surge in the diagnosis of drug-induced pancreatitis since the beginning of human immunodeficiency virus drug therapy. Didanosine has been shown to induce pancreatitis besides inducing a state of impaired glucose tolerance. Various mechanisms have been proposed, including direct toxicity, hypersensitivity, and other indirect mechanisms such as hypercalcemia and intravascular thrombosis. 202 Probably more than one mechanism operates simultaneously. Drugs such as pentamidine can cause pancreatitis weeks to months after exposure, possibly through accumulation of toxic metabolites. Iatrogenic Pancreatitis A variety of surgical procedures can lead to pancreatitis. Endoscopic retrograde cholangiopancreatography (ERCP) is associated with biochemical pancreatitis in up to 15% of cases. Acute pancreatitis following procedures performed on or near the pancreas is believed to result from direct injury to the gland or an obstruction to the flow of pancreatic juice. Such operations are exemplified by common bile duct exploration, sphincteroplasty, splenectomy, and distal gastrectomy. 203,204 The other types of surgeries that lead to pancreatitis are procedures leading to systemic disturbances that result in pancreatitis. Cardiopulmonary bypass and cardiac transplantation have been associated with pancreatitis. 205,206 Systemic hypotension leading to hypoperfusion as well as atheromatous emboli to the pancreatic circulation appears to be the mechanism in such cases. High doses of calcium in the perioperative period have also been implicated in pancreatitis following cardiopulmonary bypass. 207 Hypothermia during the procedure may also play a role because acute pancreatitis has been shown to develop as a result of exposure to low temperatures. 208 Infectious Causes of Pancreatitis Although many infectious agents, eg, mumps virus, Coxsackie virus, and mycoplasma have been implicated in the pathogenesis of pancreatitis, the exact mechanism for this association remains unclear to date. 209 Although it is generally assumed that the pathogenesis in this scenario involves direct infection of pancreatic acinar cells, no studies in humans have ever recovered the pathogen from the pancreas, although a rise in antibody titer has been observed. This rise in the antibody titer may represent a nonspecific or an anamestic response. Many parasitic infesta-