1558 WITT ET AL GASTROENTEROLOGY Vol. 132, No. 4 Clinical Features IV. A ﬁnal stage, mostly without acute ﬂares and ab- The 3 major clinical features of CP are pain, mal- sence or decreased frequency of pain, possibly asso-digestion, and diabetes. ciated with evidence of endocrine and exocrine in- sufﬁciency (burnout, see below) Pain Single stages may be skipped, eg, some patients Abdominal pain is the most vexing clinical prob- initially may present with a painless stage IV chroniclem and the most common indication for surgical inter- pancreatitis, showing maldigestion, steatorrhea, orvention in patients with CP. Severe pain decreases appe- diabetes.tite, thereby contributing to malnutrition and weightloss. The pain is usually epigastric in location (although Natural Historymore diffuse pain in the upper abdomen can occur) andmay radiate to the back. Although recurrent (type A) or The natural history of CP has been difﬁcult tocontinuous (type B) pain is considered to be the hallmark characterize because of the variability in presentationof CP, a subgroup of patients may have no pain at all, of the disease and the relative inaccessibility of thepresenting instead with symptoms of pancreatic insufﬁ- pancreas to histologic assessment. However, severalciency. While the course of pain in CP can be unpredict- studies involving large series of medical and surgicalable, in general it is reported to improve or resolve with cases have provided some important insights in thistime in the majority of patients. Whether the alleviation area.5,9,10,12–14of pain coincides with the onset of exocrine insufﬁciency Alcohol-induced CP usually develops after a prolonged period (5–15 y) of heavy alcohol consumption and does(burn-out hypothesis, see below) is still a matter of de- not develop after an isolated bout of heavy drinking. In abate.9,10 In patients with known CP, pain also may result recent report (published in 2005), Mullhaupt et al10 an-from an acute attack of pancreatitis, from a pancreatic alyzed a series of 343 patients with CP (265 patients withpseudocyst, portal or splenic vein thrombosis, or bile alcoholic CP, 57 with idiopathic CP, and 11 with hered-duct obstruction (associated with jaundice). Associated itary pancreatitis). They reported that the median age atgastric or duodenal ulcers also may contribute to pain in onset of alcoholic pancreatitis is 36 years, whereas that ofthese patients. hereditary pancreatitis was as early as 10 years. Idiopathic CP has 2 forms of clinical presentation: an early onset Maldigestion and Diabetes (juvenile) form with a median age at onset of 23 years, Steatorrhea and weight loss are further important and a late-onset (senile) form with a median age at onsetfeatures of CP. Steatorrhea is a symptom of advanced of 62 years. Tropical pancreatitis is characterized by andisease and does not occur until pancreatic lipase secre- early onset (mean age, 22 y), rapid progression, and severetion is reduced to less than 10% of normal. Maldigestion pancreatic damage in the absence of a history of alcoholof lipids occurs earlier than that of other nutrients (pro- abuse or biliary disease.4 On the other hand, autoim-teins and carbohydrates) since lipase secretion decreases mune pancreatitis is reported to occur at a later age, withmore rapidly than protease or amylase secretion. In ad- a mean age at onset of 59.4 years.8dition to exocrine insufﬁciency, diabetes mellitus may The median time to the development of pancreaticdevelop in the long-term course of the disease. The dia- insufﬁciency after disease onset depends on the type ofbetes is classiﬁed as type IIIc according to the American pancreatitis under consideration. In alcoholic and late-Diabetes Association11 and is characterized by destruc- onset idiopathic pancreatitis, exocrine insufﬁciency de-tion of both insulin- and glucagon-producing cells. The velops earlier than in early onset idiopathic pancreati-diabetic state often is fragile because the co-existing de- tis9,10,15; in alcoholic CP, pancreatic insufﬁciency canﬁciency of glucagon synthesis aggravates hypoglycemic develop as early as 6 years after the onset of disease.10situations. Similarly, endocrine insufﬁciency occurs earlier in alco- holic pancreatitis with a median time of 8 years, com- Classiﬁcation pared with 27 years in early onset idiopathic pancreatitis. Chronic pancreatitis may be separated into 4 dif- In tropical pancreatitis, both exocrine and endocrine in-ferent stages: sufﬁciency is reported to be evident at very early stages, often at the time of presentation in the majority (70%) of I. A pre-clinical stage with absent or uncharacteristic patients.4 symptoms With respect to the progression of pancreatic insufﬁ- II. Recurrent acute episodes of pancreatitis without def- ciency over time, there are conﬂicting data. Studies by inite signs of CP Lankisch et al12 and other groups16 –18 described noIII. Further recurrent episodes with intermittent or con- change or even slight improvements in pancreatic func- stant pain in between and signs of CP such as duct tion over time in patients with CP. In contrast, Mull- dilatation and pancreatic calciﬁcation on imaging haupt et al10 reported a progressive deterioration of pan-
April 2007 CHRONIC PANCREATITIS 1559 and digestive system. In their recently reported series of alcoholic and idiopathic CP patients, Mullhaupt et al10 reported that the 3 major causes of death were cardio- vascular disease, severe infection, and malignancy. Pathogenesis of Chronic Pancreatitis Research into the pathogenesis of CP was initially focused on large and small pancreatic ducts and then on the pancreatic parenchymal and nonparenchymal cells. In more recent times, the genetics of CP has attracted considerable attention and has revolutionized our knowl- edge of the possible mechanisms mediating pancreatic injury (this topic is discussed in more detail later in the section titled “Genetics of Chronic Pancreatitis”). The majority of studies related to the pathogenesis of CP haveFigure 1. Necrosis-ﬁbrosis concept of progressive pancreatic injury. focused on alcohol-induced CP. (The focus on alcohol inRepeated attacks of acute pancreatic necroinﬂammation result in in- this article reﬂects the large amount of available scientiﬁccreasing residual damage to the pancreas, eventually resulting in irre- literature on the topic. Relatively little is known aboutversible damage to the gland, characterized by acinar atrophy andﬁbrosis. the pathogenesis of acute episodes in tropical or autoim- mune pancreatitis, although there is a growing body of literature dealing with autodigestive injury in hereditarycreatic function during a median follow-up period of 16 acute pancreatitis.) This is not surprising, given thatyears in patients with alcoholic pancreatitis. The reasons alcohol abuse is the most common association of CP.for these discrepant ﬁndings are unclear, but may reﬂect Traditionally, alcoholic pancreatitis has been thought ofdifferences in study design, duration of follow-up evalu- as a form of CP from the start, punctuated during itsation, and/or differences in the sensitivities of the tests course by acute exacerbations. This notion was based onused to assess pancreatic function. studies showing that histologic and radiologic evidence The course of the pain of CP is unpredictable in indi- of CP was evident in the pancreas of many patients at thevidual patients. However, in general, pain is reported to time of their ﬁrst attack of pancreatitis.24,25 Furthermore,improve or resolve with time in the majority of patients autopsy studies had reported evidence of pancreatic ﬁ-with CP. In this regard, Mullhaupt et al10 reported that brosis in alcoholics with no clinical history of pancreati-240 of 251 patients (95.6%) with alcoholic pancreatitis tis.26 However, this concept has been challenged in recentachieved pain relief after a median time of 10 years years, with current opinion favoring the necrosis-ﬁbrosis(range, 0 –30 y) and that, in the majority of patients, this hypothesis that alcoholic pancreatitis begins as an acutepain relief coincided with the onset of exocrine and process that progresses to chronic irreversible damage asendocrine pancreatic insufﬁciency (pancreatic “burn- a result of repeated acute attacks (Figure 1).out”). However, other earlier series reported no correla- The necrosis-ﬁbrosis concept is supported by bothtion of pain relief with pancreatic insufﬁciency.9,12 Absti- clinical and experimental data. A large prospective studynence from alcohol is another important factor has reported that clinical manifestations of CP (exocrineinﬂuencing pancreatic dysfunction and pain in patients and endocrine dysfunction) were more likely to occur inwith alcoholic CP. Abstainers have a slower rate of dete- alcoholics with frequent clinical recurrent acute at-rioration of pancreatic function and a better response to tacks.10,15 In addition, a postmortem study of patientspain therapy than nonabstainers.10,15,19,20 with fatal acute alcoholic pancreatitis has shown that in The risk of developing pancreatic cancer is signiﬁcantly 53% of patients there was no evidence of chronic changeshigher in patients with CP than in the general popula- in the pancreas.27tion.21 Alcoholic CP and tropical pancreatitis are associ- Experimental evidence in support of the necrosis-ﬁbro-ated with a 15-fold and a 5-fold increased risk of pancre- sis hypothesis has accumulated rapidly in recent yearsatic cancer, respectively,21,22 whereas the cumulative and suggests that this concept is applicable not only tolifetime risk of cancer in patients with hereditary pancre- alcoholic CP but also to non–alcohol-related pancreatitisatitis is reported to be as high as 40%.21 (such as hereditary and tropical pancreatitis), in which Mortality in CP, particularly alcoholic pancreatitis, is the clinical course is punctuated with recurrent attacks ofapproximately one-third higher than that in an age- and pancreatic necroinﬂammation. Animal models of pancre-sex-matched general population.23 However, only one atic ﬁbrosis have now been developed by inducing re-ﬁfth of this excess mortality can be attributed directly to peated episodes of acute necroinﬂammation in the pan-pancreatitis itself. Most of the deaths in CP are caused by creas using an inhibitor of superoxide dismutase28 or bythe effects of alcohol and/or smoking on the liver, lungs, administration of supraphysiologic doses of cerulein
1560 WITT ET AL GASTROENTEROLOGY Vol. 132, No. 4with or without other measures such as ethanol admin- precipitation within pancreatic ducts precedes acinaristration or pancreatic duct obstruction.29,30 Most re- damage. However, recent reports of an association be-cently, Vonlaufen et al31 have demonstrated that repeated tween mutations of the cystic ﬁbrosis transmembranepancreatic necroinﬂammation induced by endotoxin ad- conductance regulator (CFTR) gene (which affect ductministration in alcohol-fed animals leads to the changes cell function) and the risk of developing idiopathic CPof CP within the gland. The molecular mechanisms re- have revived interest in the possible role of ductularsponsible for pancreatic ﬁbrosis after necroinﬂammatory dysfunction in pancreatic injury.41,42 The association be-episodes now are understood better, largely due to the tween CFTR mutations and alcoholic pancreatitis is atcharacterization of the cells that play a critical role in the present uncertain. Nonetheless, the possibility that theﬁbrogenic process, namely, the pancreatic stellate cells duct cell (in addition to the acinar cell) is an important(PSCs; see below). site of alcohol-induced injury cannot be discounted. In this regard, it is of interest to note that as early as in Alcohol-Induced Pancreatic Injury 1965, Sarles et al43 had reported that patients with alco- Studies in the ﬁeld of alcoholic pancreatitis often holic pancreatitis manifested increased levels of sweathave been hampered by the lack of suitable animal mod- electrolytes (chloride and sodium), suggesting CFTR dys-els of the disease and the difﬁculty in obtaining human function in this disease.pancreatic tissue for analysis. Nonetheless, signiﬁcant There is some evidence to suggest that chronic alcoholadvances have been made in recent years, particularly consumption facilitates protein plug formation withinwith respect to the direct toxic effects of alcohol on the pancreatic ducts. This includes: (i) increased total proteinpancreatic acinar cell, which may predispose the gland to concentration of pancreatic juice in alcoholics44; (ii) annecroinﬂammation and the role of PSCs in the produc- increased capacity of acinar cells to synthesize lithos-tion of pancreatic ﬁbrosis. tathine on alcohol exposure45 (lithostathine is a known Investigations into the pathogenesis of alcoholic pan- constituent of protein plugs with a propensity for pre-creatitis usually have followed 1 of 2 approaches, based cipitation); and (iii) an alcohol-induced decrease in acinaron 2 fundamental clinical observations. One observation content of glycoprotein GP246 (possibly because of in-is that the incidence of alcoholic pancreatitis is propor- creased secretion into pancreatic juice); this glycoproteintional to the level of alcohol consumption, suggesting has unique self-aggregating properties and is an impor-the presence of dose-related effects of alcohol on the tant constituent of protein plugs. Thus, it is possible thatpancreas.32–34 The other observation is that only a mi- blockage of small intralobular ducts by protein precipi-nority of alcoholics develop pancreatitis, suggesting that tates hinders acinar cell secretion, thereby blocking thean additional cofactor or susceptibility factor is required exit of digestive enzymes and predisposing the cell toto trigger overt disease.35,36 acute autodigestive injury (see below). Effect of alcohol on pancreatic acinar cells. Over Constant Effects of Alcohol on the Pancreas the past 3 decades, the focus of research in alcoholic Effect of alcohol on large ducts. Early research pancreatitis has shifted from pancreatic ducts to theefforts in this area (inspired by Opie’s37 observations pancreatic acinar cell itself. This focus is understandableregarding the mechanism responsible for gallstone pan- given that the cells produce large amounts of digestivecreatitis) were focused on the effects of alcohol on large enzymes (6 –20 g/day), with the potential to cause con-ducts and, in particular, the sphincter of Oddi (SO). The siderable tissue damage. The acinar cell is normally pro-large-duct theories (biliary-pancreatic reﬂux, duodeno- tected from digesting itself by synthesizing most zymo-pancreatic reﬂux, and the stimulation-obstruction the- gens as inactive precursors, by segregating zymogens intoory) postulated that altered motility of the SO in re- membrane-bound organelles, and by intracellular anti-sponse to alcohol administration played a central role in proteases. Disruption of these normal protective mecha-the development of the disease. However, unresolved nisms results in premature intracellular activation ofcontroversy about the effects of alcohol on SO function digestive enzymes, leading to autodigestive injury.and pancreatic secretion means that these theories re- Trypsinogen can be autoactivated or activated by themain of doubtful relevance to the pathogenesis of alco- lysosomal enzyme cathepsin B.47 Active trypsin, in turn,holic pancreatitis (see review by Apte et al38). can activate other pro-enzymes and trigger a digestive Effect of alcohol on small ducts. In the 1970s, enzyme activation cascade within the cell.researchers shifted their focus from large to small pan- Evidence to support a role for digestive enzymes increatic ducts, mainly as a result of the work of Sarles,39,40 pancreatitis comes from several in vitro and in vivo stud-who proposed that alcoholic pancreatitis was caused by ies (see Apte et al48 for a review), but perhaps the mostthe precipitation of secreted pancreatic proteins within compelling evidence to date in support of this theory hassmall pancreatic ducts, leading to acinar atrophy and been provided by the identiﬁcation of mutations in theﬁbrosis. The protein plug theory often has been ques- cationic trypsinogen gene in patients with hereditarytioned because of the lack of clear evidence that protein pancreatitis. As detailed later in the section on Genetics
April 2007 CHRONIC PANCREATITIS 1561of Chronic Pancreatitis, 2 mutations in particular thase) have been identiﬁed in the pancreas. Furthermore,(R122H and N29I) are known to be gain-of-function oxidant stress has been shown to occur in both humanmutations, resulting in the synthesis of an altered form and rat pancreas after ethanol exposure, most likely be-of trypsin that is resistant to inactivation.2,3,49 A role for cause of increased production of reactive oxygen speciesdigestive enzymes also has been invoked for pancreatitis (known by-products of ethanol oxidation) and decreasedrelated to hypercalcemia (via trypsinogen activation and antioxidant defenses.trypsin stabilization) and tobacco smoking (via reducedtrypsin inhibitory capacity).3 Effect of Toxic Metabolites of Ethanol Acetaldehyde, FAEEs, and reactive oxygen species Effect of Alcohol on Pancreatic Enzymes all have been shown to cause deleterious effects on the Several studies have indicated that chronic alco- pancreatic acinar cell (see review by Apte et al38). Acetal-hol administration produces changes in the acinar cell, dehyde causes morphologic damage to both rat and dogwhich may favor premature activation of digestive en- pancreas and also has been reported to inhibit stimulatedzymes. Apte et al50 have shown that messenger RNA secretion from isolated pancreatic acini. Oxidant stress(mRNA) levels and protein content of the digestive en- may contribute to the destabilization of zymogen gran-zymes trypsinogen, chymotrypsinogen, and lipase, as well ules and lysosomes observed in ethanol-fed rats (notedas the lysosomal enzyme cathepsin B, is increased in the earlier). FAEEs also have been shown to damage thepancreas of alcohol-fed rats. This increase in enzyme pancreas and its subcellular organelles. As reviewed com-content is accompanied by an increase in the fragility of prehensively by Apte et al,58 infusion of FAEEs in ratsthe organelles that contain these enzymes (zymogen leads to pancreatic edema, acinar vacuolization, andgranules and lysosomes, respectively).51,52 The effect of trypsinogen activation, and to increased extracellular ma-alcohol on lysosomal fragility is thought to be mediated trix protein levels (a ﬁnding that may have relevance toby cholesteryl esters and fatty-acid ethyl esters53,54 (sub- the development of alcohol-induced pancreatic ﬁbrosis).stances known to accumulate in the pancreas after Some of the intracellular signaling molecules that maychronic alcohol consumption55,56). The mechanism re- play a role in ethanol-induced acinar cell toxicity havesponsible for the alcohol-induced increase in zymogen now been identiﬁed. Gukovskaya et al59 have shown thatgranule fragility is unclear, but a study by Apte et al46 ethanol, acetaldehyde, and FAEEs modulate the levels ofsuggests that it may be a consequence of reduced GP2 transcription factors nuclear factor B and activator pro-levels in zymogen granule membranes since this glyco- tein-1 in parenchymal (acinar) cells, which in turn regu-protein is known to determine the shape and stability of late the expression of cytokines that mediate pancreaticzymogen granules. Alcohol-induced oxidant stress may necroinﬂammation. More recently, FAEEs also have beenbe another factor that plays a role in lysosomal and shown to cause a sustained increase in the second mes-zymogen granule membrane destabilization (see below). senger Ca within acinar cells, an effect that is thoughtThe net effect of the alcohol-induced increase in digestive to result in mitochondrial depolarization and celland lysosomal enzyme content in the presence of de- death.60creased stability of the corresponding organelles wouldbe an increased likelihood of contact between lysosomal Effect of Alcohol on Pancreaticand digestive enzymes, thereby leading to premature in- Microcirculationtracellular activation of digestive enzymes and autodiges- An aspect of pancreatic physiology that until re-tive injury to the gland. cently largely had been ignored with respect to the patho- genesis of alcoholic pancreatitis is the microcirculation Alcohol Metabolism by Pancreatic Acinar of the gland. However, 2 recent studies have shown that Cells acute on chronic ethanol administration to rats signiﬁ- Taking their cues from studies of ethanol-induced cantly decreased pancreatic perfusion.61,62 This effect wasliver toxicity, researchers postulated that ethanol may be associated with an increase in leukocyte adhesion andmetabolized by the pancreatic acinar cell to generate increased expression of adhesion molecules and cyto-toxic metabolites that may mediate the changes in the kines in the pancreas. These studies suggest that ethanol-subcellular organelles described above. Indeed, in vitro induced disturbances in pancreatic microcirculation maystudies with cultured acinar cells and isolated acini have contribute to the processes of pancreatic injury, but fur-now shown convincingly that the pancreas metabolizes ther work is required to conﬁrm and characterize theseethanol via both the oxidative and nonoxidative path- effects.ways, generating the metabolites acetaldehyde and fattyacid ethyl esters (FAEEs), respectively (see review by Wil- Individual Susceptibility to Alcoholicson and Apte57). Enzymes catalyzing ethanol oxidation Pancreatitis(alcohol dehydrogenase, cytochrome P4502E1, and cata- As seen from the preceding discussion, it is clearlase) and nonoxidative ethanol metabolism (FAEE syn- that alcohol exerts direct, constant, and toxic effects on
1562 WITT ET AL GASTROENTEROLOGY Vol. 132, No. 4the pancreas that predispose the gland to autodigestion Progression of Acute Pancreatitis to Chronicand necroinﬂammation. This is most likely the case in all Pancreatitispersons who drink heavily. However, as alluded to earlier, As noted earlier, it now is generally accepted thatit also is clear that only a minority of heavy drinkers the development of CP is the result of progressive (ac-develop acute pancreatitis, indicating that an additional crued) pancreatic damage after recurrent episodes of pan-insult or second hit is required to precipitate a clinical creatic necroinﬂammation. A few years ago, Schneiderattack of pancreatitis. The search for this trigger factor/ and Whitcomb69 proposed the sentinel acute pancreatitiscofactor/susceptibility factor has prompted numerous event hypothesis to explain the progression to CP. Theystudies over the past 2 decades, with a number of possible postulated that the “sentinel” event in this disease is acandidate factors scrutinized. These have included diet, bout of acute pancreatic injury, which makes the glandamount and type of alcohol consumed, the pattern of particularly vulnerable, in the recovery phase, to addi-alcohol consumption, hyperlipidemia, and smoking (see tional insults such as alcohol, metabolic stress, and oxi-Haber et al63 for review). As discussed in a recent review, dative stress.the role of smoking in alcoholic pancreatitis is particu- Research efforts toward elucidating the molecularlarly controversial.64 Several inherited factors have also mechanisms of CP, particularly pancreatic ﬁbrosis, werebeen studied (as discussed later in the section on Genet- given signiﬁcant impetus with the identiﬁcation, isola-ics of Chronic Pancreatitis). Therefore, it is somewhat tion, and characterization of stellate cells in the pancreasdisappointing that despite the extensive search, the fac- (reviewed by Apte et al38,70). PSCs are similar morpholog-tor(s) that unequivocally confer(s) increased susceptibil- ically to hepatic stellate cells, the principal effector cellsity to alcoholic pancreatitis remain(s) unknown. There in liver ﬁbrosis.71 It is now established that activatedremain candidate factors that have not yet been exam- PSCs play a key role in the ﬁbrogenic process in CP viained fully, including polymorphisms of proteins relevant their ability to regulate both the synthesis and degrada-to cellular antioxidant defenses and polymorphisms of tion of the extracellular matrix proteins that comprisealcohol-metabolizing enzymes, particularly FAEE syn- ﬁbrous tissue.38,70thases, minor CF mutations, and environmental factors Evidence from both clinical and experimental studiessuch as bacterial endotoxin. indicates a role for PSCs in ethanol-induced pancreatic Experimentally, putative triggers that have been exam- ﬁbrosis (see Apte et al38 for review). In vivo studies ofined for alcoholic pancreatitis include the secretagogue tissue from human beings with alcoholic pancreatitis andcholecystokinin (CCK) and bacterial endotoxin. There is from animals with experimental pancreatic ﬁbrosis havesome evidence that prior alcohol administration sensi- shown the presence of activated PSCs in areas of ﬁbrosis.tizes rodent pancreas to injury by supraphysiologic levels In vitro studies have established that PSCs are activatedof CCK,30,65 but the clinical relevance of CCK as a trigger directly by ethanol and acetaldehyde as assessed by in-factor has to be questioned. In human beings CCK is creased extracellular matrix (ECM) protein production byreleased only in picomolar quantities after meals, there- the cells. Of particular interest is the observation that ratfore it is difﬁcult to envisage a situation in which abnor- PSCs show alcohol dehydrogenase activity, indicatingmally high levels of CCK would be released into the that, apart from parenchymal (acinar) cells, ethanol alsocirculation to trigger pancreatitis in alcoholics. can be metabolized by nonparenchymal cells in the pan- In contrast to CCK, endotoxin represents a more plau- creas. Activation of PSCs by ethanol can be completelysible, physiologically relevant, trigger factor for alcoholic inhibited by the alcohol dehydrogenase (ADH) inhibitorpancreatitis. This is because: (i) increased gut permeabil- 4-methylpyrazole, indicating that ethanol-induced PSCity with translocation of gram-negative bacteria (such as activation likely is mediated by its oxidative metabolite,Escherichia coli) across the mucosal barrier is known to acetaldehyde. Furthermore, both ethanol and acetalde-occur after chronic alcohol intake in both human beings hyde cause oxidant stress within cultured PSCs and,and experimental animals66,67; (ii) plasma lipopolysaccha- importantly, incubation of PSCs with ethanol or acetal-ride (LPS, an endotoxin that is a component of bacterial dehyde in the presence of the antioxidant vitamin Ecell walls) levels have been shown to be signiﬁcantly prevents the activation of PSCs by the 2 compounds.higher in drinkers (either after chronic alcohol intake or These ﬁndings suggest that ethanol-induced PSC activa-a single binge) compared with nondrinkers68; and (iii) tion is most likely mediated by its metabolism (via ADH)endotoxemia is known to be predictive of the severity of to acetaldehyde, and the subsequent generation of oxi-acute pancreatitis (regardless of cause). It is of interest, dant stress within the cells. Interestingly, the observa-therefore, that a recent study by Vonlaufen et al31 showed tions by Vonlaufen et al31 of pancreatic ﬁbrosis in alco-signiﬁcant pancreatic necroinﬂammation in alcohol-fed hol-fed rats challenged with LPS are strongly supportedrats injected with 1 dose of LPS, and more importantly, by the in vitro ﬁndings of a synergistic effect of alcoholthe development of progressive injury as evidenced by and LPS on PSC activation.pancreatic ﬁbrosis in alcohol-fed rats challenged with During prolonged heavy alcohol intake, PSCs could berepeated doses of LPS. exposed not only to ethanol and its metabolites and LPS,
April 2007 CHRONIC PANCREATITIS 1563but also to proinﬂammatory cytokines released during Genetics of Chronic Pancreatitisepisodes of ethanol-induced pancreatic necroinﬂamma- More than 50 years ago, it was recognized for thetion. Cytokines such as tumor necrosis factor , interleu- ﬁrst time that CP may cluster in selected families, sug-kins 1 and 6, monocyte chemotactic protein, transform- gesting an inherited disease in these patients.75 The un-ing growth factor , platelet-derived growth factor derlying genetic defect, however, remained obscure for(known to be up-regulated during acute pancreatitis) more than 4 decades. As stated in this ﬁrst report oneach have been reported to activate PSCs in vitro.38 Of inherited pancreatitis, “hereditary chronic relapsing pan-particular note is that PSCs are themselves capable of creatitis does not present earmarks which distinguish itsynthesizing cytokines, and endogenous cytokine pro- from nonhereditary chronic relapsing pancreatitis.”75 Induction by the cells is stimulated by factors such as 10%–30% of patients suffering from CP, no apparentethanol, acetaldehyde, and other cytokines.72,73 These ob- underlying cause, including heredity, can be identiﬁed.servations suggest that, in addition to paracrine path- Recent research indicates that a signiﬁcant percentage ofways of activation, PSCs also may be activated in an these patients with so-called idiopathic CP may also have aautocrine manner (via endogenous cytokines), which genetic basis for their condition. The section below de-could cause perpetuation of cell activation, even when lineates the different genes involved in the pathogenesisthe initial trigger factors are no longer present. Such of hereditary or idiopathic pancreatitis, the impact ofpersistent PSC activation may potentiate ECM produc- these genetic discoveries on other types of CP such astion by the cells, eventually causing pancreatic ﬁbrosis. alcohol-related CP and tropical calciﬁc pancreatitis, and the implications for disease pathogenesis. From the above, it is apparent that during chronicalcohol consumption, PSCs are likely to be activated by 2 Cationic Trypsinogen (PRSS1)pathways operative in vivo—the necroinﬂammatory path- In 1896, Chiari76 postulated that pancreatitis re-way (via cytokines) and the nonnecroinﬂammatory path- sults from autodigestion of the gland. An inappropriateway (direct effects of ethanol and its metabolites and conversion of pancreatic zymogens to active enzymesoxidant stress). The identiﬁcation of a nonnecroinﬂam- within the pancreatic parenchyma was proposed to ini-matory pathway of stellate cell activation implies that tiate the inﬂammatory process. A key role has been at-tissue necrosis or inﬂammation may not be an absolute tributed to the activation of trypsinogen to trypsin, con-prerequisite for the stimulation of ﬁbrogenesis in the verting all proteolytic proenzymes to their active form.pancreas during alcohol abuse. Three different trypsinogens have been described in human pancreatic juice and have been designated, ac- Pathogenesis of Autoimmune Pancreatitis cording to their electrophoretic mobility, as cationic Autoimmune pancreatitis (AIP) is a relatively un- trypsinogen (PRSS1), anionic trypsinogen (PRSS2), andcommon, non–alcohol-related form of CP that has re- mesotrypsinogen (PRSS3). Compared with the anionicceived increasing attention in recent years. Only about isoenzyme, the cationic trypsinogen autoactivates more150 cases worldwide have been reported to date (the easily and is more resistant to autolysis.majority being from Japan).7 Autoimmune pancreatitis is By linkage analysis, several groups located a gene forcharacterized by the presence of (i) increased serum gam- hereditary pancreatitis on the long arm of chromosomemaglobulin levels (particularly IgG4); (ii) the presence of 7 (7q35). Subsequently, a mutation in the cationicautoantibodies (antinuclear antibodies, antilactoferrin trypsinogen gene, also referred to as serine protease 1 (PRSS1) (OMIM 276000), was identiﬁed as 1 of severalantibodies, anticarbonic anhydrase antibodies, and rheu- possible underlying defects. In 5 families, a c.365G Amatoid factor); (iii) pancreatic ﬁbrosis with lymphocytic transition leading to a substitution of arginine by histi-inﬁltration and an absence of pancreatic calciﬁcation; (iv) dine at residue 122 (p.R122H) segregated with the dis-an association with other autoimmune diseases; and (v) ease.77 R122H appears to be the most common PRSS1response to steroid therapy. The majority view of pancre- mutation observed worldwide. Subsequent studies haveatologists is that AIP does not present as acute attacks. reported other PRSS1 alterations including p.A16V,However, Takayama et al74 reported that a third of AIP p.N29I, p.N29T, p.R116C, and p.R122C, as well as severalpatients on prednisolone therapy in their series suffered others, in families with suspected hereditary pancreatitisfrom recurrent attacks of acute pancreatitis over a me- or in patients without a family history (for detailed in-dian follow-up period of 54.5 months. The pathogenesis formation of the different variants see: www.uni-leipzig.of this disease remains largely unknown but from clinical de/pancreasmutation). The functional relevance ofand experimental studies it is postulated that aberrant PRSS1 mutations has been examined by studies usingHLA-DR expression (in AIP HLA-DR expression has been recombinant cationic trypsinogen subjected to site-di-found on pancreatic ductal and acinar cells) leads to the rected mutagenesis. Mutations such as N29I and R122Hpresentation of autoantigens to lymphocytes, resulting in enhance trypsinogen autoactivation; R122H also inhibitsan autoimmune response. autolysis of the active enzyme.78,79 Thus, gain-of-function
1564 WITT ET AL GASTROENTEROLOGY Vol. 132, No. 4mutations leading to enhanced intrapancreatic trypsino- creatic trypsin activity, thereby playing a protective rolegen activation may be the common initiating step of against CP. Although the overall contribution of G191Rpancreatitis caused by PRSS1 mutations, whereas stabili- to disease pathogenesis is low, the functional character-zation of trypsin may be an accessory mechanism. ization of G191R provides the ﬁrst example in pancreati- Two PRSS1 variants, p.E79K and p.A16V, display tis for a disease-protective genetic variant.unique features: E79K trypsinogen does not alter cata-lytic activity or autolysis of trypsin, nor does it inﬂuence Serine Protease Inhibitor, Kazal Type 1inhibition of its activity by the trypsin inhibitor (SPINK1)SPINK1. However, it activates anionic trypsinogen, The serine protease inhibitor, Kazal type 1PRSS2, at least 2-fold better than wild-type cationic (SPINK1) (OMIM 167790), also known as pancreatic secre-trypsin. Thus, E79K can lead to increased trypsinogen tory trypsin inhibitor, is thought to be a potent inhibitor ofactivation by transactivation of PRSS2 instead of au- intrapancreatic trypsin activity. SPINK1 was ﬁrst isolatedtoactivation,80 but its pathogenic relevance remains to in the bovine pancreas by Kazal et al in 1948.86 It pos-be elucidated because this variant also has been found sesses a reactive site that serves as a speciﬁc target forin healthy controls. Recombinant A16V also has no trypsin. However, trypsin inhibition by SPINK1 is onlyeffect on trypsinogen activation. Instead, it increases temporary because the trypsin-SPINK1 complex itself(by 4-fold) the rate of activation peptide processing serves as a substrate for trypsin, resulting in the subse-mediated by chymotrypsin C, resulting in accelerated quent degradation of the inhibitor molecule and resto-trypsinogen activation in vitro.81 In contrast to R122H ration of the original trypsin activity.87and N29I, which display a penetrance of 70%– 80%, The focus on SPINK1 mutations as possible pathoge-A16V is found almost exclusively in patients without a netic factors in pancreatitis was the result of the knowl-family history of pancreatitis.82 Recently, a triplication edge that a signiﬁcant number of hereditary pancreatitisof an approximately 605-kb segment containing PRSS1 patients do not show a PRSS1 mutation, indicating thatand PRSS2 was reported in 5 families with hereditary defects in other genes might be involved in disease patho-pancreatitis.83 Thus, besides point mutations, a gain of genesis. It was hypothesized that, in addition to gain-of-trypsin through a gene-dosage effect also may contrib- function mutations in PRSS1 as a cause of pancreatitis,ute to the disease pathogenesis. The importance of CP also may be a result of “loss-of-function” mutationsPRSS1 mutations as pathogenic mediators in heredi- in pancreatic trypsin inhibitors. A mutation in thetary pancreatitis is supported by a recent study using a SPINK1 gene (a c.101A G transition leading to substi-transgenic mouse model expressing mutant R122H tution of asparagine by serine at codon 34 [p.N34S]) hasmouse trypsinogen. The pancreas of these mice dis- been found in 18 of 96 unrelated pediatric pancreatitisplayed early onset acinar injury, inﬂammatory cell in- patients; 6 patients were homozygous for this muta-ﬁltration, and enhanced response to cerulein-induced tion.88 No phenotypic differences between heterozygouspancreatitis. With progressing age, pancreatic ﬁbrosis and homozygous N34S patients were detected. This as-and acinar cell dedifferentiation developed.84 sociation between N34S and CP has now been conﬁrmed by several other studies. Anionic Trypsinogen (PRSS2) N34S is found mostly in patients without a family Because increased proteolytic activity caused by history of CP: 15%– 40% of patients with so-called idio-mutated PRSS1 enhances the risk for CP, it was thought pathic CP carry N34S on 1 allele or on both alleles. Datathat mutations in the anionic isoenzyme PRSS2 (OMIM from 8 large studies in Europe and the United States601564) also may predispose to disease. Notably, how- indicate that 12.6% of patients with CP are heterozygousever, a recent study indicated that the PRSS2 mutation and 3.6% are homozygous for N34S, whereas only 1.9% ofmay be a protective factor against CP. A recent study controls are heterozygous for this variant. Interestingly,reported a c.571G A transition resulting in substitution the N34S mutation also has been reported in about halfof glycine by arginine at codon 191 (p.G191R), which was the patients with tropical calciﬁc pancreatitis from In-found in 220 of 6459 (3.4%) controls but only in 32 of dia.89 The pathogenic action of N34S, however, remains2466 (1.3%) patients (odds ratio, 0.37; P 1.1 10-8).85 elusive. Recombinant N34S mutated human SPINK1Further analyses showed that patients (with hereditary, does not show any altered trypsin inhibitor capacity.90 Itidiopathic, and alcoholic pancreatitis) with G191R were is worth noting that N34S is in complete linkage disequi-of an older age than those without the protective variant. librium (LD) with 4 other intronic sequence variants: In vitro studies showed that recombinant G191R pro- c.56-37T C, c.87 268A G, c.195-604G A, and c195-tein, on activation by enterokinase or trypsin, showed a 66_-65insTTTT.88 Thus, it may be speculated that it iscomplete loss of trypsin activity due to the introduction one of these intronic alterations and not N34S itself thatof a novel tryptic cleavage site that renders the enzyme is the pathogenic relevant mutation.hypersensitive to autocatalytic proteolysis. Thus, it ap- The second most common SPINK1 mutation,pears that the G191R PRSS2 variant mitigates intrapan- c.194 2T C, affects position 2 of the splice donor site
April 2007 CHRONIC PANCREATITIS 1565in the third intron, which is highly conserved in eu- was expected and in idiopathic CP it was 4 times askaryotes. Analysis of mutated mRNA shows a truncated expected. In the other study, 17 CFTR mutations in 27SPINK1 because exon 3 is skipped.91 Several other patients with idiopathic CP were investigated.41 SevenSPINK1 alterations have been described in recent years, patients (25.9%) had at least 1 CFTR mutation and 1mainly in single patients or families only (for detailed patient was compound heterozygous. The frequency ofinformation of the different variants see: www.uni- CFTR mutations in idiopathic CP was 6 times higherleipzig.de/pancreasmutation). With the exception of a than expected. However, both these studies only investi-few mutations that strongly suggest a loss of function by gated the most common of the approximately 1000 CFTRdestruction of the ATG initiation codon (c.2T C) or by mutations that now have been described. Subsequentshift of reading frame with premature termination studies analyzing the complete CFTR coding sequence as(c.27delC, c.98dupA), the functional consequences of well as PRSS1 and SPINK1 in idiopathic CP patientsmost variants are unknown. Recently, expression studies found that 25%–30% of patients carried at least 1 CFTRof 2 dominant inherited mutations affecting the signal mutation, but that only a few patients were compoundpeptide, c.41T C (p.L14P) and c.41T G (p.L14R), re- heterozygous.98,99 Thus, idiopathic CP actually may rep-ported a rapid intracellular degradation of the mutant resent “atypical” cystic ﬁbrosis caused by the combina-inhibitor molecules leading to abolished SPINK1 secre- tion of 2 mild or of 1 mild and 1 severe CFTR mutation.tion.92 Similar to and perhaps even more pronounced Several CP patients, however, were transheterozygous forthan for PRSS1, SPINK1 mutations display a marked a CFTR alteration and a SPINK1 or PRSS1 variant, illu-variability of penetrance and inheritance pattern. Some minating the signiﬁcance of the combination of muta-variants that are likely to lead to complete functional loss tions in different genes in disease pathogenesis.98,99of the mutated allele such as c.2T C, c.27delC, or codon14 mutations, appear to follow a dominant trait, whereas Alcoholic Pancreatitisthe N34S variant appears to be a recessive or complex The association between alcohol abuse and pan-trait. creatitis is well established, but individual susceptibility The role of SPINK1 in pancreatitis has been evaluated to alcohol varies widely and only a minority of heavyrecently in 2 genetically engineered animal models. drinkers develop CP. Increasing evidence portends thatTransgenic expression of rat Spink1 in mice, which leads additional environmental or genetic cofactors are neces-to an increased endogenous trypsin inhibitor capacity by sary, which are mostly unknown. Several studies investi-190%, signiﬁcantly reduced the severity of cerulein-in- gating PRSS1, pancreatitis associated protein, 1-antit-duced pancreatitis,93 while targeted disruption of Spink3 rypsin, CFTR, cytokeratin 8, major histocompatibility(the murine homologue of human SPINK1) resulted in complex antigens, and alcohol metabolizing or detoxify-autophagic degeneration of acinar cells, impaired regen- ing enzymes have yielded negative or conﬂicting results.eration, and death within 2 weeks after birth.94 In the Since xenobiotic-mediated cellular injury is thought tolatter model, enhanced tryptic activity was detected in play a role in the pathogenesis of alcoholic CP, geneticpancreatic acini prepared 1 day after birth.95 variations reducing the activity of detoxifying biotrans- formation enzymes have also been examined. Recently, a Cystic Fibrosis Transmembrane Conductance low detoxiﬁcation activity allele of the UDP-glucurono- Regulator (CFTR) syltransferase 1A7, UGT1A7*3, was linked to pancreatic Cystic ﬁbrosis transmembrane conductance regu- cancer and alcoholic CP.100 However, a subsequent studylator (CFTR) is an apical membrane chloride channel could not conﬁrm these ﬁndings.101critical for ﬂuid and electrolyte secretion in the respira- In a large multicenter study, an association betweentory and digestive tracts. In the pancreas, CFTR is local- mutated SPINK1 and alcoholic CP was described: theized to centroacinar and proximal ducts and regulates N34S mutation was found in 16 of 274 (5.8%) patientsductal bicarbonate secretion. Abnormal CFTR function with alcoholic CP, but only in 4 of 540 (0.8%) healthyas a result of CFTR gene mutations is associated with CF, control individuals and 1 of 98 (1.0%) alcoholic controlsan autosomal-recessive disease characterized by pulmo- without CP.102 Subsequent studies have reported annary dysfunction and pancreatic insufﬁciency. A minority N34S frequency in alcohol-related CP of around 6%.of CF patients suffer from recurrent pancreatitis. In 1998, Most recently, the protective PRSS2 variant, G191R (see2 studies described an association between idiopathic CP below), was reported to be signiﬁcantly less common inand mutations in the CFTR gene (OMIM 602421).96,97 patients with alcoholic CP compared with healthy con-One study tested 134 patients with CP, including 60 trols (5 of 609 [0.8%] vs 220 of 6459 [3.4%]; P .0001).85patients with idiopathic and 71 patients with alcohol- In summary, recent advances in the ﬁeld of genetics ofinduced disease, for 22 mutations.97 Eighteen patients CP have substantially improved our understanding of the(13.4%), including 12 with idiopathic CP (20%), were disease (Figure 2). For a long time, hereditary pancreatitisheterozygous for a CFTR mutation. The frequency of was thought to be a rare disorder. However, the ﬁndingsCFTR mutations in alcohol-related CP was twice what of PRSS1, SPINK1, and CFTR mutations in patients with
1566 WITT ET AL GASTROENTEROLOGY Vol. 132, No. 4Figure 2. Role of digestive enzymes in pancreatitis. (A) Normal pancreas. Trypsin resulting from autoactivation of trypsinogen within the pancreaticparenchyma is inhibited by SPINK1 and also by mesotrypsin or trypsin (as a second-line defense). This defense mechanism prevents the pancreasfrom activating the pancreatic enzyme cascade and autodigestion. (B) Pancreatitis. Mutations in PRSS1 or in SPINK1 lead to an imbalance ofproteases and their inhibitors within the pancreatic parenchyma, resulting in an inappropriate activation of pancreatic zymogens with subsequentautodigestion and inﬂammation. Mutations in CFTR also may disturb the delicate balance between proteases and antiproteases, by intrapancreaticacidiﬁcation or by defective apical trafﬁcking of zymogen granules, thereby facilitating the intrapancreatic activation of digestive enzymes. Dark boxesrepresent products of mutated genes. AP, activation peptide. Modiﬁed from Witt et al.85so-called idiopathic CP indicate that cases of inherited CP ate trigger factor (environmental or genetic) then stimu-are much more common than originally envisioned. lates overt pancreatic necrosis. Repeated episodes ofThese data challenge the differentiation between “hered- acute necroinﬂammation (regardless of etiology) and theitary” and “idiopathic” pancreatitis. Different mutations release of proinﬂammatory cytokines leads to the activa-in different genes might lead to different phenotypic tion of pancreatic stellate cells (PSCs). PSCs also arepresentations and inheritance patterns, and even the activated directly by ethanol (via its metabolite acetalde-same mutation in the same gene might have different hyde and the subsequent generation of oxidant stress).consequences depending on the individual’s genetic Persistent activation of PSCs leads to an imbalance be-background and environmental factors. The discovery of tween extracellular matrix protein synthesis and degra-SPINK1 mutations in other types of CP such as tropical dation, eventually resulting in pancreatic ﬁbrosis, a car-calciﬁc pancreatitis and alcohol-induced CP further blurs dinal feature of CP.the borders between the particular CP subtypes. It isanticipated that the identiﬁcation of other genes involved Diagnosis of Chronic Pancreatitisin the pathogenesis of inherited CP will also enhance our The diagnosis of CP relies on relevant symp-knowledge about more common types of CP such as toms, imaging modalities to assess pancreatic struc-alcoholic or tropical CP. Future research most likely will ture, and assessment of pancreatic function. The diag-reveal a very complex interaction between various envi- nostic gold standard of early stage disease would be anronmental and genetic factors, with ﬂowing transitions adequate surgical biopsy, which is rarely available.among these subtypes (Figure 3). However, because the primary lesions of early stage CP Current Concept of the Pathogenesis of are usually focal, ﬁne-needle biopsy examinations may Chronic Pancreatitis The clinical and experimental evidence (largelybased on alcohol-related studies) has led to the followingconcept for the pathogenesis of CP (Figure 4). The pancreas may be rendered susceptible to autodi-gestive injury, either because of abnormal trypsin activa-tion/inactivation mechanisms (as in hereditary, meta-bolic, and tropical pancreatitis) or because of the effectsof toxins such as ethanol (via its metabolites and its Figure 3. Inﬂuence of different environmental and genetic factors onmetabolic by-products) on digestive and lysosomal en- the pathogenesis of chronic pancreatitis. ACP, alcoholic chronic pancre-zyme content within the acinar cell and on the stability of atitis; TCP, tropical calciﬁc pancreatitis; ICP, idiopathic pancreatitis; HP,the organelles that contain these enzymes. An appropri- hereditary pancreatitis.
April 2007 CHRONIC PANCREATITIS 1567Figure 4. Current concept of the pathogenesis of chronic pancreatitis. Three major elements of the pancreas are implicated in the development ofirreversible pancreatic damage. (1) An acinar cell that is susceptible to autodigestive injury for the following reasons (depending on cause): (a) theeffects of ethanol and its metabolites on subcellular organelles including increased digestive and lysosomal enzyme content (secondary to increasedsynthesis [increased mRNA] and impaired secretion) and destabilization of lysosomes and zymogen granules; (b) impairment of trypsinogenactivation/deactivation processes. In the presence of an appropriate trigger factor, overt acinar cell injury is initiated. (2) A pancreatic stellate cell thatis activated by cytokines released during pancreatic necroinﬂammation or by direct effects of ethanol, its metabolites, and oxidant stress, leading toexcessive extracellular matrix protein production. (3) A pancreatic ductule blocked by protein precipitation, which may further facilitate diseaseprogression. AC, acetaldehyde; CE, cholesteryl esters; L, lysosome; ZG, zymogen granule.yield false-negative results. In the absence of deﬁnite Serum Parameterssigns of CP, it often is difﬁcult to differentiate early In patients presenting with pain, levels of pancre-stage disease from recurrent acute pancreatitis. For a atic enzymes (mainly lipase) are determined in order todeﬁnitive diagnosis, a careful history and follow-up identify an acute episode of the disease. In patients with-may be necessary. It is important that the assessment out acute attacks, reduced serum trypsinogen, lipase, orof a patient with painful CP includes upper gastroin- amylase levels may be found; however, the sensitivities oftestinal endoscopy, abdominal ultrasonography, and these tests are less than 60%, so none of them per se areendoscopic retrograde cholangiopancreatography helpful in diagnosing CP.104(ERCP) or magnetic resonance cholangiopancreatogra-phy (MRCP) in order to detect a potentially reversible Imaging Procedurescause of pain (eg, peptic ulcer, pseudocyst, commonbile duct stricture). Several imaging procedures are available for the The correlation between structural and functional im- evaluation of CP. Transabdominal ultrasound andpairment of the pancreas in CP is often poor.103 Patients computerized tomography are the most commonlywith severe exocrine insufﬁciency may have a largely used techniques, whereas endoscopic retrogradenormal pancreatic structure and vice versa. In general, cholangiopancreatography (ERCP), endoscopic ultra-advanced stages of CP may be diagnosed easily by imag- sonography (EUS), and MRCP usually are restricted toing procedures such as ultrasound, computerized tomog- specialized centers.raphy (CT), magnetic resonance tomography, and mag-netic resonance cholangiopancreatography (MRCP). In Abdominal X-raycontrast, the diagnosis of early disease presents a consid- Plain abdominal x-ray has lost its place in theerable challenge. modern diagnostic imaging paradigm. Nonetheless, in The biochemical, structural, and functional parame- some cases, calciﬁcation on an abdominal x-ray whenters used to assist in the diagnosis of CP are discussed associated with steatorrhea can clinch the diagnosis andbelow: obviate the need for a further, more extensive, work-up.
1568 WITT ET AL GASTROENTEROLOGY Vol. 132, No. 4 Transabdominal Ultrasonography At present, it is still unclear what criteria may be used to Transabdominal ultrasonography is an inexpen- diagnose mild CP reliably in patients with a normalsive technique representing the ﬁrst procedure usually ERCP. MRCP is regarded as useful in patients at high riskperformed in patients with suspected CP. Duct alter- of developing post-ERCP pancreatitis, with a low proba-ations, calciﬁcation, and cysts and are detected with high bility of ductal alterations, or inaccessibility of the pan-sensitivity by this modality. Other complications of pan- creatic duct as a result of pancreatic or gastric surgery.111creatitis such as duodenal or gastric distention and bile Diagnosis of Pancreatic Cancer in Chronicduct dilatation also may be demonstrated. In patients Pancreatitiswith excessive abdominal gas or acute pancreatitis asso- As noted earlier, patients with CP have an in-ciated with ileus, the view is often limited, making the creased risk of developing pancreatic cancer.21 Currently,procedure highly related to the investigator’s skills. Nev- no imaging procedure can reliably detect a malignantertheless, sonography is a simple technique and, in the tumor reliably in patients with CP. A recent study re-hands of experienced investigators, remains a useful ported that 2-(18)F)-ﬂuoro-2-deoxy-D-glucose positronmethod for rapid and reliable diagnosis. emission tomography can differentiate between neoplas- Endoscopic Retrograde tic and inﬂammatory pancreatic tumors at a sensitivity of Cholangiopancreatography (ERCP) 91% and a speciﬁcity of 87%.112 However, whether 2-(18)F)-ﬂuoro-2-deoxy-D-glucose positron emission to- ERCP is regarded as the gold standard for the mography is superior to traditional imaging techniquesdetection of CP. Typical alterations of pancreatic ducts such as CT or magnetic resonance imaging (MRI) isobserved on ERCP are dilatations, stenoses, and abnor- questionable because the latter also have been reported tomalities of the side branches. The duct structure also may classify CP and pancreatic cancer correctly in 90% ofbe used to stage the disease according to the Cambridge patients.classiﬁcation.105 However, Cambridge stage I is oftenquestioned as a reliable ﬁnding indicating CP. The most Functional Studiesimportant role of ERCP is the identiﬁcation of structural The secretin-cerulein test is regarded as the “goldabnormalities such as duct stenosis, stones, or cysts that standard” for the detection of exocrine pancreatic insuf-may be amenable to interventional treatment and, the ﬁciency.113 However, the procedure is only available in aexclusion, if possible, of pancreatic cancer. It is to be few specialized centers and its protocol as well as itsnoted that ERCP eventually may be superseded by a evaluation is not well standardized. In addition, the testnoninvasive alternative, namely MRCP, for the diagnosis is time consuming and uncomfortable for the patient.of CP. Therefore, less invasive alternatives have been developed Endoscopic Ultrasonography (EUS) and including fecal elastase, lipase, or chymotrypsin; the pan- Magnetic Resonance creolauryl test; the bentiromide test; and a variety of Cholangiopancreaticography (MRCP) breath tests using radiolabeled pancreatic substrates, usually triolein. However, none of these tests have been The role of endosonography for diagnosing early able to meet clinical needs unequivocally. In mild orstage CP is not well deﬁned. The technique is regarded as moderate pancreatic insufﬁciency, the sensitivity of thesethe most sensitive procedure to detect the disease. Thir- tests is inadequate. It is only in severe disease that pan-teen criteria, such as reduced or increased echogenicity, creatic function tests show a high sensitivity114; however,increased lobulation, and alteration of small and large the diagnosis of severe CP is usually obvious by otherducts, have been described.106 It is accepted that the means, making a pancreatic function test unnecessary. Inabsence of these criteria reliably rules out CP, whereas the patients with mild or moderate disease, pancreatic func-presence of 5 or 6 criteria strongly indicates the diagno- tion tests only achieve sensitivities of 50% and 65%–75%,sis.107 The signiﬁcance of less than 5 criteria, however, is respectively, and hence they are not very helpful in theunclear. Apart from interobserver variability, one has to diagnostic work-up of patients with recurrent pain oftake into account the fact that a nonhomogeneous echo unclear origin. Other MRI– derived tests have been de-structure is not a speciﬁc sign for CP but also can be seen scribed, but the majority of these studies did not assessin the normal pancreas, especially in the elderly.108 There patients with mild or moderate CP.115,116have been 2 follow-up studies of patients with suspectedCP with a normal ERCP, but with alterations on EUS.One study showed no alterations on ERCP during a Treatment of Chronic Pancreatitisfollow-up period of 12–38 months,109 whereas the other The treatment of CP is mainly symptomatic and issuggested a rapid progression of disease because ductal directed toward the cardinal features of pain, and exo-changes were observed on ERCP after a mean follow-up crine and endocrine insufﬁciency. A diagnosis of CP doesperiod of 18 months in 22 of 38 patients.110 The latter not necessarily require treatment because patients may beﬁnding indicates a surprisingly rapid progression to CP. asymptomatic. However, if a precipitating factor such as
April 2007 CHRONIC PANCREATITIS 1569an anatomic anomaly or a metabolic disease can be iden- mately 25,000 –50,000 U lipase/meal are recommended,tiﬁed, it may be treated by surgical or medical interven- but a higher dose or combination with a proton pumption. In general, the therapeutic strategies for CP include inhibitor may be required.abstinence from alcohol and cigarette consumption, painrelief, correction of exocrine and endocrine insufﬁciency, Treatment of Diabetesnutritional support, and endoscopic or surgical interven- Diabetes in CP is classiﬁed as type IIIc.11 However,tion. the treatment is not different from patients with type I In alcoholic patients, the major goals of treatment are diabetes. Due to the co-existing deﬁciency of glucagon,sustained abstinence from alcohol and smoking, the im- patients with CP have an increased risk of hypoglycemicprovement of compliance, and social re-integration. Most events. This is a particular problem in patients with pooralcoholic patients are heavy smokers and, clinically, it can compliance and/or continued alcohol consumption orbe difﬁcult to achieve abstinence from both alcohol and autonomic neuropathy. In these patients, the therapeutictobacco. In a practical sense, the treating physician may goal should be to avoid hypoglycemia by a simple insulinhave to concede continued smoking as a trade-off for regimen. As indicated earlier, the survival of patients withalcohol abstinence. Although the role of alcohol absti- alcoholic CP is limited. Approximately 50% of patientsnence in the reduction of pancreatic pain is somewhat will not live longer than 10 years after the initial diagno-unclear, there is evidence to indicate that deterioration of sis and therefore will not beneﬁt from aggressive insulinpancreatic function is slower in abstainers than in non- therapy. A more intensive insulin regimen is indicatedabstainers,9,20 and that abstainers have a better response only in patients with good compliance and cessation ofto pain therapy than nonabstainers. The role of smoking alcohol. Acarbose and insulin sensitizers are ineffective.as an causative factor in CP is controversial.64,117,118Nonetheless, cigarette consumption contributes to the Nutritionexcess mortality associated with the condition. There is no such thing as a speciﬁc pancreatic diet. Pain Abstinence from alcohol and the intake of smaller but In CP, abdominal pain is a serious clinical prob- more frequent meals is recommended. Restriction of fatlem leading to a markedly compromised quality of life intake is not advised if the pancreatic exocrine insufﬁ-and, potentially, narcotic addiction. Treatment of pain ciency is largely compensated by enzyme-replacementshould be started with conventional analgesics such as therapy. Restriction of dietary fat and administration ofacetaminophen. If pain relief is not achieved, additional medium-chain triglycerides (MCTs) is indicated only inprescription of opiates may be necessary. However, it is cases of severe maldigestion refractory to treatment sinceimportant to be mindful of the well-known side effects of MCTs may worsen diarrhea in many patients. Deﬁcien-opioids such as central nervous system depression, alter- cies of fat-soluble vitamins are found mainly in patientsations of gastrointestinal motility, and induction of de- who continue to drink; in these cases vitamin supplemen-pendence. Other, as yet unproven, strategies of pain relief tation can be instituted.include inhibition of pancreatic enzyme secretion usingpancreatic enzyme therapy and the use of antioxidants. Interventional Treatment of ComplicationsInvasive approaches such as celiac plexus block, endo- There is a long-standing controversy concerningscopic procedures, and surgical drainage and resection the indication(s) for interventional (mainly endoscopic)also have been used as therapy for the pain of CP, but therapy of complications of CP such as duct stones,none of these procedures has ever been the subject of pancreatic or biliary tract stenosis, or pseudocysts. Therecontrolled trials either in comparison with medical ther- has been a profound lack of randomized controlled stud-apy or with no therapy. ies, resulting in therapeutic decisions largely being made on the basis of technical skills available rather than Treatment of Exocrine Insufﬁciency scientiﬁc evidence. Many centers perform interventional In theory, pancreatic enzymes are indicated in therapy only in symptomatic patients with recurrent painpatients with steatorrhea (fecal fat 7 g/day) and weight or acute attacks, associated with ductal dilatation proxi-loss. In clinical practice, however, measurement of fecal mal to the stenosis or an obstructive stone. Some evi-fat rarely is performed, so that the decision for enzyme dence exists that stenting a biliary stenosis is inferior toreplacement is based on an assessment of the patient’s surgery; transitional stenting has been shown to be effec-clinical state. The dose of pancreatic enzymes given tive only in patients with a mass in the head of theshould be high enough to treat steatorrhea, but a signif- pancreas obstructing the pancreatic duct.119 Interven-icant increase of body weight is rarely achieved. These tional therapy is obligatory in patients with cyst-associ-enzymes should be taken with meals in acid-protected ated pain, gastric compression, or biliary obstruction(enteric-coated) formulations (except in patients after when alternatives are absent (eg, if surgery is refused bygastric surgery, ie, Kausch–Whipple resection). Approxi- the patient or the surgeon).
1570 WITT ET AL GASTROENTEROLOGY Vol. 132, No. 4 Surgery 4. Balakrishnan V, Nair P, Radhakrishnan L, Narayanan VA. Tropi- cal pancreatitis—a distinct entity, or merely a type of chronic Indications for surgery include complications pancreatitis? Indian J Gastroenterol 2006;25:74 – 81.such as common bile duct or duodenal obstruction, fail- 5. Ammann RW, Akovbiantz A, Largiader F, Schueler G. Course andure of endoscopic therapy in a patient with intractable outcome of chronic pancreatitis. Longitudinal study of a mixedpain, or a suspected pancreatic cancer.120 Like endoscopic medical-surgical series of 245 patients. Gastroenterology 1984; 86:820 – 828.therapy, surgical procedures for pain in CP have never 6. Kloppel G, Maillet B. [Development of chronic pancreatitis frombeen subjected to randomized controlled trials compar- acute pancreatitis: a pathogenetic concept.] Zentralbl Chiring them with medical therapy or no therapy. In patients 1995;120:274 –277.with pain and an inﬂammatory tumor of the head of the 7. Ketikoglou I, Moulakakis A. Autoimmune pancreatitis. Dig Liverpancreas, a duodenum-preserving resection of the head is Dis 2005;37:211–215. 8. Okazaki K. Autoimmune-related pancreatitis. Curr Treat Optionsthe method of choice in many centers.121 As an alterna- Gastroenterol 2001;4:369 –375.tive procedure, a longitudinal prancreaticojejunostomy 9. Layer P, Yamamoto H, Kalthoff L, Clain JE, Bakken LJ, DiMagnomay be considered if the main pancreatic duct is dilated EP. The different courses of early- and late-onset idiopathic andto 7 mm or more.122 alcoholic chronic pancreatitis. Gastroenterology 1994;107: 1481–1487. 10. Mullhaupt B, Truninger K, Ammann R. Impact of etiology on the Summary painful early stage of chronic pancreatitis: a long-term prospec- In summary, CP is characterized by progressive tive study. Z Gastroenterol 2005;43:1293–1301.and ultimately irreversible pancreatic injury that mani- 11. The Expert Committee on the Diagnosis and Classiﬁcation of Diabetes Mellitus. Follow-up report on the diagnosis of diabetesfests clinically as maldigestion and diabetes. Alcohol mellitus. Diabetes Care 2003;26:3160 –3167.abuse is the most common association of CP in the 12. Lankisch PG, Lohr-Happe A, Otto J, Creutzfeldt W. NaturalWestern world. Important advances have been made in course in chronic pancreatitis. Pain, exocrine and endocrinerecent years with respect to our understanding of the pancreatic insufﬁciency and prognosis of the disease. Digestionpathogenesis of this disease, particularly related to the 1993;54:148 –155.mechanisms responsible for the development of pancre- 13. Lankisch PG, Seidensticker F, Lohr HA, Otto J, Creutzfeldt W. The course of pain is the same in alcohol- and nonalcohol-atic ﬁbrosis (a cardinal feature of CP) after repeated acute induced chronic pancreatitis. Pancreas 1995;10:338 –341.attacks of pancreatic necroinﬂammation (the necrosis- 14. Lowenfels AB, Zwemer FL, Jhangiani S, Pitchumoni CS. Pancre-ﬁbrosis concept). The pancreatic stellate cell is now es- atitis in a native American Indian population. Pancreas 1987;tablished as playing a central role in ﬁbrogenesis, partic- 2:694 – 697.ularly when activated either directly by toxic factors 15. Ammann RW, Muellhaupt B. Progression of alcoholic acute to chronic pancreatitis. Gut 1994;35:552–556.associated with pancreatitis (such as ethanol, its metab- 16. Garcia-Puges AM, Navarro S, Ros E, et al. Reversibility of exo-olites, or oxidant stress) or by cytokines released during crine pancreatic failure in chronic pancreatitis. Gastroenterologypancreatic necroinﬂammation. Considerable research ef- 1986;91:17–24.fort also has been directed toward the genetic abnormal- 17. Kondo T, Kayakawa T, Noda A, et al. Follow-up study of chronicities that may predispose to CP. Mutations of several pancreatitis. Gastroenterology Japan 1981;16:46 –53. 18. Miyake H, Harada H, Kunichika K, Ochi K, Kimura I. Clinicalcandidate genes related to trypsinogen activation/inacti- course and prognosis of chronic pancreatitis. Pancreas 1987;vation and to CFTR function increasingly are being rec- 2:378 –385.ognized for their potential disease-modiﬁer role in dis- 19. Ammann RW, Heitz PU, Kloppel G. Course of alcoholic chronictinct forms of CP including alcoholic, tropical, and pancreatitis: a prospective clinicomorphological long-termidiopathic pancreatitis. Treatment of uncomplicated CP study. Gastroenterology 1996;111:224 –231. 20. Gullo L, Barbara L, Labo G. Effect of cessation of alcohol use onis usually conservative, with the major aim being to the course of pancreatic dysfunction in alcoholic pancreatitis.effectively alleviate pain, maldigestion, and diabetes, and, Gastroenterology 1988;95:1063–1068.consequently, to improve the patient’s quality of life. 21. Lowenfels AB, Maisonneuve P. Risk factors for pancreatic can-Surgical and endoscopic interventions are reserved for cer. J Cell Biochem 2005;95:649 – 656.complications such as pseudocysts, abscesses, and malig- 22. Chari ST, Mohan V, Pitchumoni CS, Viswanathan M, Madanago-nancies. palan N, Lowenfels AB. Risk of pancreatic carcinoma in tropical calcifying pancreatitis: an epidemiologic study. Pancreas 1994; 9:62– 66. References 23. Levy P, Milan C, Pignon JP, Baetz A, Bernades P. Mortality 1. DiMagno EP, Layer P, Clain JE. Chronic pancreatitis. In: Go VLW, factors associated with chronic pancreatitis. Gastroenterology DiMagno EP, Gardner JD, Lebenthal L, Reber HA, Scheele GA, 1989;96:1165–1172. editors. The pancreas: biology, pathobiology and disease. New 24. Howard JM, Ehrlich EW. The etiology of pancreatitis. A review of York: Plenum Press, 1993:665–706. clinical experience. Ann Surg 1960;152:135–137. 2. Ahmed SA, Wray C, Rilo HL, et al. Chronic pancreatitis: recent 25. Strum WB, Spiro HM. Chronic pancreatitis. Ann Intern Med advances and ongoing challenges. Curr Probl Surg 2006;43: 1971;74:264 –272. 127–238. 26. Pitchumoni CS, Glasser M, Saran RM, Panchacharam P, Thelmo 3. Etemad B, Whitcomb DC. Chronic pancreatitis: diagnosis, clas- W. Pancreatic ﬁbrosis in chronic alcoholics and nonalcoholics siﬁcation, and new genetic developments. Gastroenterology without clinical pancreatitis. Am J Gastroenterol 1984;79:382– 2001;120:682–707. 388.