Gut: first published as 10.1136/gut.29.10.1413 on 1 October 1988. Downloaded from

Gut, 1988, 29, 1413-1423

Progress report Heretical thoughts on the pathogenesis of acute

Acute haemorrhagic pancreatitis was first described by R H Fitz 99 years ago, and, seven years later, H Chiari suggested that the pancreatic necrosis was caused by a protein autodigestion. l2 Fitz had believed that the pancreatitis originated 'by extension of a [bacterial] gastroduodenal inflam- mation along the pancreatic duct', but, on the report of the bacteriologist, W H Welch, on two of his patients, and from reports on five others in the literature, E L Opie found that, because of the variable and inconstant results of bacterial culture, the various organisms were not aetiological agents but merely opportunistic invaders of injured tissues.?5 Opie thought that the disease was due to reflux ofbile into the pancreatic duct, writing that 'the peculiar necrosis which affects the pancreatic parenchyma is doubtless referable to trypsin which is contained in the pancreatic cells'. His common channel theory has now been discarded, but four other of his assumptions, which all derived from those same seven patients, have seldom been questioned and have become, of course, accepted truths.6 Opie's four postulates were, first, that acute haemorrhagic pancreatitis (HP) results from autodigestion by pancreatic proteolytic enzymes (to which http://gut.bmj.com/ may be added the question of how and where the enzymes are activated); second, bacterial plays no part in the pathogenesis of such pancreatitis; third, acute interstitial pancreatitis (IP) is a mild, and sometimes early, form of haemorrhagic pancreatitis and not an entity in itself; and, fourth, (AP) is caused by obstruction to the outflow from the duct system, rather than by reflux into it. pancreatic on September 26, 2021 by guest. Protected copyright. These assumptions are now examined, but, in doing so, multifactorial and other aetiological aspects of AP will hardly be mentioned. The chief emphasis will be on those patients with HP, rather than on the great majority who have a self-limiting disease and soon get well. As Williamson has put it: 'Of any four patients with acute pancreatitis, three will rapidly improve with simple supportive therapy, while the fourth develops one or more serious complications of the disease and stands about a one-in-three chance of dying.'7 Necrosis of haemorrhagic pancreatitis is not caused by proteolytic autodigestion Autodigestion by pancreatic proteolytic enzymes has been assumed to explain the necrosis of HP, despite the paucity of evidence. Over 80 years, most work on AP has concentrated on trypsin and the activation of trypsinogen: far less attention has been devoted to the other pancreatic proteases, chymotrypsin, elastase and carboxypeptidase, and their pre- cursors normally activated in the by enterokinase and 1413 Gut: first published as 10.1136/gut.29.10.1413 on 1 October 1988. Downloaded from

1414 Keynes trypsin. For activation to occur in the pancreas, the obvious mechanism would be by reflux of enterokinase and trypsin from the , though activation of the precursors in pancreatic juice is complicated by the presence of other enzymes, the pH, divalent cations and inhibitors. This would be in the duct system, however, and not explain intracellular activation. Steer, Meldosi, and Figarella found that lysosomal hydrolases could activate the pancreatic proteases intracellularly, but then could only suggest that this subsequently caused acinar cell injury.8 Other activators have been sought, particularly in bile, ever since Opie's common channel theory was promulgated: uninfected bile does not activate trypsinogen.9 Intraductal injection of bile will give AP if under unphysio- logical pressure, as will injection of many fluids, including distilled water. '1" But, it is the toxic properties of bile salts, and not the pancreatic proteases, that causes the inflammatory response. 12 The main pathological feature of HP is parenchymatous necrosis, but activated pancreatic proteases have not been shown to be cytotoxic.12-14 Intraperitoneal injection of actively proteolytic pancreatic juice is not toxic if bacteria free, and injection of trypsin into the lobular spaces of the canine pancreas gives pancreatic oedema with interlobular haemorrhage, but not necrosis.'215 The injection into the pancreas alters the microcirculation, producing vasodilatation, increased capillary permeability, and local oedema with very little alteration in the parenchymal architecture. Similar injection of carboxypeptidase, or amylase, produces little or no reaction, but elastase damages the microcirculation. '5 Intradermal injection of sterile trypsin gives an oedematous wheal without cell necrosis.16

The lytic actions of the pancreatic proteases are limited and specific. http://gut.bmj.com/ Trypsin does not attack the phospholipids of cell membranes, though the hydrophilic, but not the hydrophobic, portions of cell surface proteins might be vulnerable to it with possible stripping of receptors. Tryptic digestion is, of course, used in tissue and cell culture to cleave cell surface attachments, suspend cells, and digest away intercellular material, connective tissue and blood clot, but not done to harm cells. The concentration of trypsin employed is between 0 1 and 0 5 %, certainly far higher than physiological, on September 26, 2021 by guest. Protected copyright. or even pathological, levels ever reached in vivo.'7 There is also indirect evidence against autodigestion by trypsin.'8 9 In AP, the similar levels of release of amylase, lipase, trypsinogen, and the three other protease precursors, are indicative of injury to, or increased pressure in, the pancreatic duct system.'8202' Once released, the enzymes may, or may not, be involved in the further disease processes of AP. In the presence of naturally occurring inhibitors, any activation of the precursors has been difficult to explain, or, for long, to show, as an immunoassay does not distinguish between a protease, its precursor, and their complexes with inhibitors. The lack of success in the treatment of AP by antiproteases can more easily be explained, it autodigestion by pancreatic proteases does not occur. Phospholipase and lysolecithin, as well as amylase, may cause fat necrosis; lipase, trypsin, and elastase do not.2223 Phospholipase, which is a prominent component of human pancreatic juice, and normally activated by trypsin and bile salts in the small intestine, reacts with phospholipid, though not that of intact cellular membranes, to produce lysolecithin. It converts the lecithin and cephalin of cell membranes and bile into their cytotoxic Gut: first published as 10.1136/gut.29.10.1413 on 1 October 1988. Downloaded from

Heretical thoughts on thepathogenesis ofacutepancreatitis 1415 lysocompounds, and it is these that can cause extensive vascular and parenchymatous necrosis.2324 A system not involving theoretical proteolytic autodigestion, but including phospholipase, could produce HP by duo- denopancreatic reflux of lysolecithin formed in the upper digestive tract. In patients with HP and IP, serum phospholipase concentrations were signific- antly higher in the former, and very high levels were recorded in those who died.24 It is possible to extend the usual meaning of 'autodigestion' to include the action of activated pancreatic phospholipase via lysolecithin. But the often reiterated statement that pancreatic necrosis results from autodigestion of the parenchyma consequent on the activation of the 'resident' proteolytic enzymes is untenable. The vast majority of research into AP has, of course, for the whole of this century involved the pancreatic proteolytic enzymes. Bacteria as a cause of acute pancreatitis Early this century, infection was widely accepted as a cause of AP: in 1901, S Flexner had found that intraductal injection of Bacillus coli, B pyocyaneus and B diphtheriae caused the 'hemorrhagic inflammation' of acute pancreatitis, and, since, it has been shown that infected bile was more likely to produce pancreatitis than if it were sterile." '82528 Armstrong, Taylor, and Torrance recently injected the pancreatic ducts of rats with sterile bile under low pressure, and with infected bile after passing it through a filter, without producing HP, though it was produced on injecting bile infected with E coli.29 But, as a result of Opie's book in 1903, there has been strikingly little research into the possibility that bacterial infection might play a part in the aetiology of AP. There has been a lack of bacterial data from preconceived http://gut.bmj.com/ theories that HP is purely due to chemical digestion, though the aseptic nature of necrotic pancreatitis is far from proven.` The Pfeffer closed duodenal loop has been widely used since 1957 for studying AP, but recently it has been called an 'inappropriate model' that should be abandoned because infection occurs as well as AP.32-3 Nance and Cain, using the preparation, however, obtained HP in germ free dogs, so that bacteria are not necessarily a primary factor in the pathogenesis of on September 26, 2021 by guest. Protected copyright. reflux HP.35 In 87 dogs given blind loops, 67% developed HP and 10% IP, but only when reflux occurred: in six dogs with HP, there were similar high bacterial counts in the loop fluid and the pancreas 24 hours after operation, but, in four with IP, fluid and pancreas were sterile. 18 Intraductal injection of infected loop fluid gave HP, but, when filtered, the fluid gave IP. Neomycin instilled into a blind loop prevented the onset of HP, and in those with HP, but not with IP, there were high concentrations of plasma amidase, concentrations that were lowered with neomycin (but not with intravenous aprotinin). Ix 3 There was a correlation between pancreatic infection, HP and raised levels of amidase. 1, The proteolytic enzyme amidase showed a broad specificity and stability uncharacteristic of pancreatic proteases, but similar to that of the bacterial enzyme preparation pronase which can cause intradermal necrosis.'8 37 In 1956, it was noticed that not all bacteria can give HP on intraductal injection, but only those that produce toxins that can damage cell membrane phospholipids and give an intradermal necrotising lesion with small vessel occlusion.3' The production of experimental HP by bacteria is dependent on Gut: first published as 10.1136/gut.29.10.1413 on 1 October 1988. Downloaded from

1416 Keynes organisms, such as clostridia, coliform, and staphylococci, that produce cytotoxins, and is preventable by immunisation, antitoxins, or anti- biotics.3'" `6 It is of interest in this context that a patient with AP from a primary campylobacter infection did not have the haemorrhagic form of the disease.4' Campylobacter coli/jejuni is not a cytotoxin producer. Clinically, the organisms grown from pancreatic , or necrotic pancreas, are invariably cytotoxin producers.42 In Beger and colleagues' first series of patients with HP, comprising 13% of all those admitted with AP (excluding 18 found to have a pancreatic ), all 114 underwent abdominal exploration.47 Microorganisms (usually E coli) were identified in necrotic pancreatic tissue from 45, or 39.4%; 21 of the operations took place within seven days of the onset of illness, and, as early as this, five, or 24%, were found to be bacteriologically positive. The disease had a far higher mortality rate in those who showed pancreatic than in those without such infection: 61% of those discovered to be infected within 21 days of onset died, but only 5.9% of those with sterile necrosis.47 In an earlier study, this time using percutaneous aspiration, there was a similar pancreatic infection rate of 40% .48 Acute pancreatitis may follow endoscopic retrograde cholangiopancrea- tography (ERCP), and in some patients the pancreatitis has been associated with infection introduced into the pancreatic duct by the endoscope.49 Gregg found that bacteria, usually Gram negative organisms, were present in the pancreatic juice of 33% of patients with AP, and that in those with infected juice the clinical course was particularly severe.' This supports the view that the pancreatic infection derives from reflux from the duodenum (as in the blind loop experiments). Infected bile might also be a cause of HP, and one report found that 14-4% of patients with biliary pancreatitis also had acute http://gut.bmj.com/ cholangitis, or 23.3% with gall stone related cholangitis showed AP.5' No adequate clinical trial has been done to evaluate the use of in AP, though many physicians routinely give them to those severely ill. Their apparent failure to affect the course of the disease has strengthened the assumption that infection is a secondary event, but the difficulties of such a study are similar to those done on the use of antiproteases. Patients with HP have a high mortality rate despite all treatment, whereas those with IP on September 26, 2021 by guest. Protected copyright. are likely to survive with only supportive therapy, and unless there is knowledge of the proportion of each type, a trial is unlikely to provide helpful results. There are similar problems with the use of antibiotics in the treatment of acute cholangitis.5253 The clinical courses of haemorrhagic pancreatitis and those of septic shock, toxic shock syndrome, fulminant , gram negative septicaemia and endotoxaemia show similarities.'4 Also, an infective HP could be the 'motor' of the multiple organ failure syndrome where bacteria, bacterial toxins, and cytotoxic effluents from inflammatory cells can play prominent parts.55 Bacterial cytotoxins could release inflam- matory mediators from monocytes (cytokines), tumour necrosis factor (cachectin), interleukins, prostaglandins, or oxygen derived free radicals, and thereby initiate acute haemorrhagic pancreatitis.is59 It has been found that oxygen derived free radicals, such as superoxide, hydrogen peroxide and the hydroxyl radical, are highly reactive toxic substances, which may be produced during oxidative metabolism by the enzyme xanthine oxidase, and can play an important role in the pathogenesis of experimental Gut: first published as 10.1136/gut.29.10.1413 on 1 October 1988. Downloaded from

Heretical thoughts on thepathogenesis ofacutepancreatitis 1417 pancreatitis.5'5 It is possible, then, that AP could, on occasion, be induced by bacterial products in the duodenopancreatic reflux fluid, even if it is not an infectious disease in the traditional sense. Necrotic pancreatitis gives similar haemodynamic changes to those observed in sepsis, perhaps from circulating vasoactive compounds.` Severe AP is commonly associated with the adult respiratory distress syndrome (ARDS), as is severe Gram negative sepsis. A recent study has suggested that the pulmonary endothelial damage in ARDS may be solely mediated by endotoxins, and the administration of endotoxins under certain conditions may cause HP in rats.6`62 Endotoxins are lipopolysaccharides found in the outermost part of the cellular membrane of Gram negative bacteria, and have been found in patients with AP, with the severity of the attack correlating with the degree of endotoxaemia.'-' All these clinical observa- tions can only suggest, and cannot be interpretated as evidence of, a primary bacterial role in the development of severe acute pancreatitis. Since this review was completed, Beger and his colleagues have reported on a further 95 patients with HP, in all of whom necrotic tissue was excised and examined bacteriologically. Infection, practically always by Gram negative 'gut' organisms and, perhaps, therefore derived from duodeno- pancreatic reflux fluid, was found in 42%.67 Those with infected necrosis more often developed organ complications, and had a far higher mortality rate, than those with sterile necrosis. The operation took place with a median duration from the onset of the illness of seven days, that is, the infection with necrosis was usually severe enough to be detected in the earliest days of the illness. Though not in any way conclusive, this could argue for a primary infection rather than rapid, contaminating, secondary

colonisation. But does a secondary infection take place all that commonly? http://gut.bmj.com/ On combining both reports of Beger and his colleagues, a total of 121 patients, found on exploration to have a sterile necrosis, were most unlikely to show a later, secondary infection, and, as already pointed out, had a far better survival than those with infected necrotic tissue.4767 This review with its emphasis on HP implies that severe AP should no longer be considered a non-infective disease and can be caused by primary infection. Opie's assumption that bacteria play no part in the pathogenesis on September 26, 2021 by guest. Protected copyright. of HP has been challenged, and it is possible that, in at least 40% of those most severely ill, it is bacteria capable of producing cytotoxins that cause the pancreatic necrosis. The bacteriology of AP requires urgent further investigation. Interstitial pancreatitis: not a mild form of haemorrhagic pancreatitits There is a clear distinction between clinical 'interstitial', or 'oedematous', and 'haemorrhagic', or 'necrotic', pancreatitis - there is no necrosis of acinar cells in IP." Haemorrhagic pancreatitis shows diffuse parenchymatous necrosis with haemorrhage secondary to the necrosis of the walls of blood vessels. The term necrotic pancreatitis might be more appropriate, but the gross appearance of the pancreas is dominated by haemorrhage. The simple finding of blood on the surface of the gland does not warrant the diagnosis of HP, and has, perhaps, led to inaccurate assessment of the severity of the disease, both clinically and experimentally. In a recent paper on HP, nine patients of 81, or 11%, did not show parenchymal necrosis on histology.`9 Gut: first published as 10.1136/gut.29.10.1413 on 1 October 1988. Downloaded from

1418 Keynes In clinical and experimental HP, the inflammatory infiltration is slight, if present at all, and typically only seen at the periphery of the gland at the interface between necrotic areas and viable tissue - any inflammatory cells are necrosed along with the acinar cells and blood vessels.'8` Clinically, three stages have been described: (a) the gland is enlarged with solitary, focal areas of necrosis, and the capsule is preserved; (b) confluent necrosis with necrosis of lipid tissue, but the capsule is preserved; (c) necrosis extends from the pancreas into the surrounding areas, especially into the omental bursa and the retroperitoneum, and the capsule is destroyed.7` Fat necrosis may be scanty. The evaluation of necrosis in early pancreatitis should be made on microscopy, where so often it has to be made biochemically. In interstitial pancreatitis, the pancreas is firm and swollen, with scattered fat necrosis. On histology, there is a patchy oedema of the interstitial tissue that is accompanied by inflammatory infiltration.68 The inflammatory response in IP usually results from leakage of fluid into the gland on damage to, or increased pressure in, the pancreatic duct system: intraductal injection of fluid under physiological, low pressure does not give AP, or a raised plasma amylase concentration, but, with injection under higher pressure, the fluid infiltrates the gland, giving interstitial oedema with an inflam- matory response, and a raised plasma amylase concentration.'0"7' Experi- mentally, Armstrong, Taylor, and Torrance found that on injection there was no extravasation at low intraduct pressures; that, at moderate pressures, there was passage of fluid through the intercellular clefts into the periacinar spaces; and, at high pressures, duct ruptures occurred.7' With pancreatic trauma without damage to the duct system there is no rise in plasma amylase; in both clinical and experimental pancreatitis, there is little correlation between the degree of elevation of amylase and the severity of http://gut.bmj.com/ any pancreatitis; plasma amylase is elevated when the pancreatic ducts are ligated or obstructed, when substances are injected into them, or on ERCP.72-75 On histological examination, the distinction between the two histo- pathological types of AP is usually clear, and HP is a separate entity from IP. Occasionally, however, the histological features may be mixed, and this, perhaps, means that an aetiological factor of HP has been added to the IP. on September 26, 2021 by guest. Protected copyright. Does clinical interstitial pancreatitis simply worsen into the haemorrhagic type if the condition deteriorates? According to Kelly, 95% of patients with AP have a self-limiting condition that subsides spontaneously, usually within 72 hours after first being seen. But, around 5% of the patients develop the necrotising form of the disease, which causes most of the early deaths, usually, from sepsis.76 Experimentally, using blind duodenal loops and obtaining reflux pancreatitis, dogs with IP showed no raised levels of the proteolytic enzyme amidase, and no bacteria were cultured from the gland. These dogs, which could survive for many days after operation, were relatively fit at 24 hours, and showed scant peritoneal fluid, but obvious fat necrosis. In dogs with HP, there were high levels of amidase, and a high bacterial count in the pancreas. The dogs were in bad condition 24 hours after operation, and showed copious peritoneal fluid, but little fat necrosis, and were unlikely to survive more than 48 hours.'8 Thus, on clinical, experimental and histopathological grounds, there is a case for considering that there are two separate forms of AP: the majority have IP and soon get well; a small minority have HP, and subsequently die, Gut: first published as 10.1136/gut.29.10.1413 on 1 October 1988. Downloaded from

Heretical thoughts on thepathogenesis ofacutepancreatitis 1419 or develop inflammatory masses. On occasion, IP may progress to HP if there is, say, secondary infection by bacteria capable of producing cyto- toxins. Change from the prognostically useful names of IP and HP to 'mild' and 'severe' AP is unlikely to be clinically beneficial. Reflux versus obstruction: the incompetent sphincter of Oddi The common belief that AP results from obstruction to the pancreatic duct was attacked by McCutcheon in 1968.77 In patients dying of AP, obstruction of the pancreatic duct is rare, and in one series was found in only three of 100 patients.78 Ligation of the duct in patients with relapsing pancreatitis does not produce further pancreatitis, and duct strictures seen in are a result, rather than a cause, of repeated inflammation.` Neoplasms of the head of the pancreas may block its ducts, but distal pancreatitis is uncommon, as it is in patients with stenosis of the sphincter of Oddi.` With acute obstruction, AP is rare: the resulting changes are like those of IP, and similar to those of leakage from the duct system into the parenchyma. The gland becomes oedematous and infiltrated by inflam- matory cells, and plasma amylase is raised for a relatively short period, which may be associated with slight fat necrosis and increased peritoneal fluid. The oedema settles and is followed by fibrosis and disappearance of the acinar tissue; necrosis is not seen.'879 If obstruction of the pancreatic duct rarely leads to pancreatitis, can reflux from the duodenum be a common cause? McCutcheon believed it was, and, experimentally, reflux from a blind loop can produce AP, which is prevented if the pancreatic duct is tied.'877 Pancreatitis is a complication of gastro- enterostomy, or a Polya partial gastrectomy, when there is obstruction of an http://gut.bmj.com/ afferent loop below the ampulla, or with obstruction of the duodenum giving, in effect, a blind loop.77 AP occurs in those whose pancreatic ducts do not have a common channel with the common , or with pancreas divisum.1'43 Conditions that interfere with the competence of the spincter of Oddi, such as carcinoma of the ampulla, may lead to pancreatitis.'0 Temporary hypotonia of the spincter of Oddi, may follow damage to the sphincter, on September 26, 2021 by guest. Protected copyright. sphincterotomy and sphincteroplasty, and AP.9-93 After 18 444 endoscopic sphincterotomies, 275 patients, or 1.5%, developed AP, from which 56, or 20-4%, died.' Kelly in 1984, and Armstrong and his colleagues the next year, found that gall stone associated pancreatitis did not occur in patients who had large gall stones completely blocking the ampulla or the pancreatic duct, but in those with smaller stones who also showed greater dilatation of the .95 9 Biliary pancreatitis is initiated by passage of a gall stone through, or lodgement at, the ampulla, causing (not necessarily persistent) hypotonia and incompetence of the sphincter, and allowing reflux.1'0' There is a high incidence of pancreatic duct reflux during operative cholangiography in patients undergoing cholecystectomy after episodes of gall stone associated AP, and in these patients the reflux of contrast accompanies a low common bile duct pressure and occurs after the entry of the dye into the duodenum.'`' 102 The reflux is also associated with duodenal segmental activity and abolished by administration of opiates.'03 There is accumulating evidence that the majority of attacks of AP are precipitated by a migrating gall stone, though often none for even a Gut: first published as 10.1136/gut.29.10.1413 on 1 October 1988. Downloaded from

1420 Keynes temporary synchronous obstruction.107-'09 The mere presence of gall stones does not establish the mechanism by which they may cause AP, here argued to follow reflux and not obstruction, but in one study, 75% of patients undergoing exploration with 48 hours of the onset of the symptoms of AP were found to have gall stones at the ampulla, but in only 25% if the operation was delayed for four to 13 days after the onset.'05 It appears likely that most of the patients who fall into the idiopathic group of those who have AP, at present perhaps as large a group as 25%, will be shown to have gall stone associated pancreatitis.'0911 " A case has been made for believing that reflux of duodenal fluid into the pancreatic duct explains the pathogenesis of AP rather than obstruction to the duct. By rendering the sphincter of Oddi incompetent, reflux can readily occur after the passage of a gall stone into the duodenum, or after partial obstruction of the ampulla, both of which are more likely if the stone is smaller. An undetected passage could account for many socalled idiopathic cases of AP, and, if the duodenopancreatic reflux contains lysolecithin, cytotoxin producing bacteria, or toxic bacterial products, the pathogenesis of acute haemorrhagic pancreatitis may be explained without mention of either proteolytic autodigestion or obstruction. MILO KEYNES Department ofAnatomy, Downing Street, Cambridge CB2 3DY.

References http://gut.bmj.com/ 1 Fitz RH. Acute pancreatitis: a consideration of pancreatic hemorrhage; hemorrhagic suppurative and gangrenous pancreatitis and disseminated fat necrosis. Boston Med Surg J 1889; 120: 181-7, 205-7, 229-35. 2 Chiari H. Ueber die Selbstverdauung des menschlichen Pankreas. Z Heilk 1896; 17: 69-96. 3 Opie EL. The etiology of acute hemorrhagic pancreatitis. Johns Hopkins Hosp Bull 1901; 12:182-8. 4 Opie EL. Disease ofthepancreas - its cause and nature. Philadelphia: JB Lippincott, 1903.

5 Opie EL, Meakins JC. Data concerning the etiology and pathology of hemorrhagic on September 26, 2021 by guest. Protected copyright. necrosis of the pancreas. J Exp Med 1909; 11: 561-78. 6 Keynes WM. The mythology of acute pancreatitis. in Surgery 1987; 6: 354-68. 7 Williamson RCN. Early assessment of severity in acute pancreatitis. Gut 1984; 25: 1331-9. 8 Steer ML, Meldolesi J, Figarella C. Pancreatitis: the role of lysosomes. Dig Dis Sci 1984; 29: 934-8. 9 Elmslie R. The effect of bile on the activation of trypsinogen and the activity of trypsin in pancreatic juice. Br J Surg 1965; 52: 465-70. 10 Rich AR, Duff GL. Experimental and pathological studies on the pathogenesis of acute haemorrhagic pancreatitis. Bull Johns Hopkins Hosp 1936; 58: 212-54. 11 Mann FC, Giordano AS. The bile factor in pancreatitis. Arch Surg 1923; 6: 1-30. 12 Dragstedt LR, Haymond HE, Ellis JC. Pathogenesis of acute pancreatitis (acute pancreatic necrosis). Arch Surg 1934; 28: 232-91. 13 Robinson TM, Dunphy JE. Continuous perfusion of bile protease activators through the pancreas. JAMA 1963; 183: 530-3. 14 Babson AL, Williams PA, Phillips GE. An evaluation of serum 'trypsin' tests. Clin Chem 1962; 8:62-6. 15 Anderson MC. Pathophysiology of acute pancreatitis. World J Surg 1981; 5: 325-7. 16 Anderson MC. Necrotizing properties of human blood following incubation with autologous pancreatic secretions. J Surg Res 1965; 5: 239-44. 17 Paul J. Cell and tissue culture. 5th ed. London: Churchill Livingstone, 1975. 18 Keynes WM. A non-pancreatic source of the proteolytic-enzyme amidase and bacteriology in experimental pancreatitis. Ann Surg 1980; 191: 187-99. Gut: first published as 10.1136/gut.29.10.1413 on 1 October 1988. Downloaded from

Hereticalthoughts on thepathogenesis ofacutepancreatitis 1421

19 Beck IT, Pinter EJ, Solymer J, et al. The role of pancreatic enzymes in the pathogenesis of acute pancreatitis. II. The fate of pancreatic proteolytic enzymes in the course of acute pancreatitis. 1962; 43: 60-70. 20 Benini L, Tagliaro F, Fioretta A, et al. Hyperenzymemia and molecular size distribution of trypsin and elastase-1 in the serum after ERCP and during acute pancreatitis [Abstract]. Gut 1985; 26: A649-50. 21 Flamion B, Delhaye M, Horanyi Z, et al. Comparison of elastase-l with amylase, lipase and trypsin-like immunoreactivity in the diagnosis of acute pancreatitis. Am J Gastro- enterol 1987; 82: 532-5. 22 Panabokke RG. An experimental study of fat necrosis. J Path Bact 1958; 75: 319-31. 23 Schmidt H, Creutzfeldt W. The possible role of phospholipase A in the pathogenesis of acute pancreatitis. Scand J Gastroenterol 1969; 4: 39-48. 24 Puolkkainen P, Valtonen V, Paananen A, Schroder T. C-reactive protein (CRP) and serum phospholipase A2 in the assessment of the severity of acute pancreatitis. Gut 1987; 28: 764-71. 25 Ivy AC, Gibbs GE. Pancreatitis: a review. Surgery 1952; 31: 614-42. 26 Flexner S. Experimental pancreatitis. Johns Hopkins Hosp Reports 1901; 9: 743-71. 27 Archbald E. The experimental production of pancreatitis in animals as the result of the resistance of the common duct sphincter. Surg Gynecol Obstet 1919; 28: 529-45. 28 Armstrong CP, Taylor TV, Torrance HB. Bile and bacteria in acute pancreatitis. Br J Surg 1984; 71: 1001. 29 Thal AP, Perry JF Jr, Egner W. A clinical and morphologic study of 42 cases of fatal acute pancreatitis. Surg Gynecol Obstet 1957; 105: 191-202. 30 Thal AP, Molestina JE. Studies on pancreatitis. III. Fulminating hemorrhagic pancreatic necrosis produced by means of staphylococcal toxin. Arch Pathol 1955; 60: 212-20. 31 Thal AP, Tansathithaya P, Egner W. An experimental study of bacterial pancreatitis. Surg Gynecol Obstet 1956; 103: 459-68. 32 Pfeffer RB, Stasior 0, Hinton JW. The clinical picture of the sequential development of acute hemorrhagic pancreatitis in the dog. Surg Forum 1957; 8: 248-51. 33 McMahon MJ. Acute pancreatitis. Curr Opin Gastroenterol 1987; 3; 698-713. 34 Dickson AP, Foulis AK, Imrie CW. Histology and bacteriology of closed duodenal loop models of experimental acute pancreatitis in the rat. Digestion 1986; 34: 15-21. 35 Nance FC, Cain JL. Hemorrhagic pancreatitis in germfree dogs. 1967; 18: 365-7. http://gut.bmj.com/ 36 Byrne JJ, Joison J. Bacterial regurgitation in experimental pancreatitis. Am J Surg 1964; 107: 317-20. 37 Bertelli A. Proteases and antiproteasic substances in the inflammatory response. Biochem Pharmacol Suppl 1968; 17: 229-40. 38 Carnot. Re'cherches sur les pancreatites. Thesis, Paris, 1898. 39 Tejerina-Fotheringham W. Acute pancreatitis: pathophysiology and treatment. Gastro- enterology 1948; 10: 687-96.

40 Konok GP, Thompson AG. Pancreatic ductal mucosa as a protective barrier in the on September 26, 2021 by guest. Protected copyright. pathogenesis of pancreatitis. Am J Surg 1969; 117: 18-23. 41 Gallagher P, Chadwick P, Jones DM, Turner L. Acute pancreatitis associated with campylobacter infection. Br J Surg 1981; 68: 383. 42 Saxon A, Reynolds JT, Doolas A. Management of pancreatic abscesses. Ann Surg 1981; 194: 545-52. 43 Kivilaakso E, Lempinen M, Makelainen A, et al. Pancreatic resection versus peritoneal lavation for acute fulminant pancreatitis. Ann Surg 1984; 199: 426-31. 44 Warshaw AL, Jin G. Improved survival in 45 patients with pancreatic abscess. Ann Surg 1985; 202:408-15. 45 Malangoni MA, Shallcross JC, Seiler JG,et al. Factors contributing to fatal outcome after treatment of pancreatic abscess. Ann Surg 1986; 203: 605-11. 46 Bradley EL III. Management of infected pancreatic necrosis by open drainage. Ann Surg 1987; 206: 542-8. 47 Beger HG, Bittner R, Block S, Buchler M. Bacterial contamination of pancreatic necrosis. Gastroenterology 1986; 91: 433-8. 48 Gerzof SG, Banks PA, Spechler SJ, et al. Role of guided percutaneous aspiration in early diagnosis of pancreatic sepsis. Dig Dis Sci 1984; 29: 950. 49 Elson CO, Hattori K, Blackstone MO. Polymicrobial sepsis following endoscopic retrograde cholangiopancreatography. Gastroenterology 1975; 69: 507-10. 50 Gregg JA. Detection of bacterial infection of the pancreatic ducts in patients with pancreatitis and pancreatic cancer during endoscopic retrograde cholangiopancreato- graphy. Gastroenterology 1977; 73: 1005-7. Gut: first published as 10.1136/gut.29.10.1413 on 1 October 1988. Downloaded from

1422 Keynes

51 Neoptolemos JP, Carr-Locke DL, Leese T, James D. Acute cholangitis in association with acute pancreatitis: incidence, clinical features and outcome in relation to ERCP and endoscopic sphincterotomy. BrJ Surg 1987; 74: 1103-6. 52 Thompson JE, Tompkins RK, Longmire WP. Factors in management of acute cholangitis. Ann Surg 1982; 195: 137-45. 53 Dooley JS, Hamilton-Miller JMT, Brumfitt W, Sherlock S. Antibiotics in the treatment of biliary infection. Gut 1984; 25: 988-98. 54 Wardle N. Bacteraemic and endotoxic shock. BrJ Hosp Med 1979; 21: 223-31. 55 Carrico CJ, Meakins JL, Marshall JC, et al. Multiple-organ-failure syndrome. Arch Surg 1985; 121: 196-208. 56 Cobo JC, Abraham E, Bland RD, et al. Sequential hemodynamic and oxygen transport abnormalities in patients with acute pancreatitis. Surgery 1984; 95: 324-30. 57 Sanfey H, Bulkley GB, Cameron JL. The role of oxygen-derived free radicals in the pathogenesis of acute pancreatitis. Ann Surg 1984; 200: 405-12. 58 Sanfey H, Bulkley GB, Cameron JL. The pathogenesis of acute pancreatitis. Ann Surg 1985; 201: 633-8. 59 Standfield NJ, Kakkar VV. Prostaglandins and acute pancreatitis. Br J Surg 1983; 70: 573-6. 60 Bradley EL III, Hall JR, Lutz J, et al. Hemodynamic consequences of severe pancreatitis. Ann Surg 1983; 198: 130-3. 61 Ognibene FP, Martin SE, Parker MM, et al. Adult respiratory distress syndrome in patients with severe neutropenia. N EnglJ Med 1986; 315: 547-51. 62 Shibayama Y. Pancreatic venous stasis and endotoxaemia as aetiologic factors in acute haemorrhagic pancreatitis. J Pathol 1987; 152: 177-82. 63 Keleti G, Sykora JL. Production and properties of cyanobacterial endotoxins. Appl Environ Microbiol 1982; 43: 104-9. 64 Playforth MJ, McMahon MJ. Do bacteria play a role in the pathogenesis of acute pancreatitis? Ann R Coll Surg Engl 1982; 64: 46. 65 Foulis AK, Murray WR, Galloway D, et al. Endotoxaemia and compliment activation in acute pancreatitis in man. Gut 1982; 23: 656-61.

66 Kivilaakso E, Valtnen VV, Malkamaki M, et al. Endotoxaemia and acute pancreatitis: http://gut.bmj.com/ correlation between the severity of the disease and the antienterobacterial common antigen antibody titre. Gut 1984; 25: 1065-70. 67 Beger HG, Buchler M, Bittner R, et al. Necrosectomy and postoperative local lavage in necrotizing pancreatitis. Br J Surg 1988; 75: 207-12. 68 Longnecker DS. Pathology and pathogenesis of diseases of the pancreas. Am J Pathol 1982; 107: 103-21. 69 Teerenhovi 0, Nordback I, Isolauri J. Influence of pancreatic resection on systemic complications in acute necrotizing pancreatitis. Br J Surg 1988; 75: 793-5. 70 Becker V. Pathological anatomy and pathogenesis of acute pancreatitis. World J Surg on September 26, 2021 by guest. Protected copyright. 1981; 5: 303-13. 71 Armstrong CP, Taylor TV, Torrance HB. Pressure, volume and the pancreas. Gut 1985; 26: 615-24. 72 Thistlethwaite JR, Hill RP. Serum amylase levels in experimental pancreatitis. Surgery 1952; 31: 495-501. 73 Ranson JHC, Spencer FC. The role of peritoneal lavage in severe acute pancreatitis. Ann Surg 1978; 187: 565-75. 74 Nordback I. Value of monitoring amylase activities in patients with pancreatitis. Lancet 1985; i: 1092. 75 Skude G, Wehlin L, Maruyama T, Anyama J. Hyperamylasaemia after duodenoscopy and retrograde cholangiopancreatography. Gut 1976; 17: 127-32. 76 Kelly TR. Pathological anatomy and pathogenesis of acute pancreatitis. World J Surg 1981; 5: 309-10. 77 McCutcheon AD. A fresh approach to the pathogenesis of pancreatitis. Gut 1968; 9: 296- 310. 78 Shader AE, Paxton JR. Fatal pancreatitis. AmerJ Surg 1966; 111: 369-73. 79 Senninger N, Moody FG, Coelho JCU, Vanburen DH. The role of biliary obstruction in the pathogenesis of acute pancreatitis in the opossum. Surgery 1986; 99: 688-93. 80 Bradley EL III, Clements JL. Idiopathic duodenal obstruction. Ann Surg 1981; 193: 638- 46. 81 Britt LG, Samuels AD. Pancreas divisum: is it a surgical disease? Ann Surg 1983; 197: 654- 61. Gut: first published as 10.1136/gut.29.10.1413 on 1 October 1988. Downloaded from

Heretical thoughts on thepathogenesis ofacutepancreatitis 1423

82 Warshaw AL, Richter JM, Schapiro RH. The cause and treatment of pancreatitis associated with pancreas divisum. Ann Surg 1983; 198: 443-50. 83 Russell RCG, Wong NW, Cotton PB. Accessory sphincterotomy (endoscopic and surgical) in patients with pancreas divisum. BrJ Surg 1984; 71: 954-7. 84 White SH, Nazarian NA, McEwen Smith A. Periampullary adenoma causing pancreatitis. Br Med J 1981; 283: 527. 85 Jones PA, Rayter Z, Knight MJ, etal. Juxtampullary, bile-filled duodenal duplication cyst: another surgically correctable cause of acute pancreatitis. J R Soc Med 1982; 75: 662-4. 86 Waldron R, Drumm J, McCarthy CF, et al. Surgically correctable recurring pancreatitis. BrJSurg 1985; 72:203. 87 Kalvaria I, Bornman PC, Girdwood AH, Marks IN. Periampullary cyst: a surgically remediable cause of pancreatitis. Gut 1987; 28: 358-62. 88 Robertson JFR, Imrie CW. Acute pancreatitis associated with carinoma of the ampulla of Vater. BrJ Surg 1987; 74: 395-7. 89 Belghiti J, Wind P, Bernades P, Fekete F. Acute pancreatitis associated with carcinoma of the ampulla of Vater. Br J Surg 1987; 74: 1067-8. 90 Dixon JM, Chapman RW, Berry AR. Carcinoid tumour of the ampulla of Vater presenting as acute pancreatitis. Gut 1987; 28: 1296-7. 91 Leese T, Neoptolemos JP, Carr-Locke DL. Successes, failures, early complications and their management following endoscopic sphincterotomy: results in 394 consecutive patients from a single centre. BrJ Surg 1985; 72: 215-9. 92 Vernava A, Andrus C, Herrmann VM, Kaminski DL. Pancreatitis after surgery. Arch Surg 1987; 122: 575-80. 93 Gregg JA, Carr-Locke DL. Endoscopic pancreatic and biliary manometry in pancreatic, biliary, and papillary disease, and after endoscopic sphincterotomy and surgical sphincteroplasty. Gut 1984; 25: 1247-54. 94 Norback, I, Airo 1. Post-ERCP acute necrotizing pancreatitis. Ann Chirurg Gynaecol 1988; 77:15-20. 95 Kelly TR. Gallstone pancreatitis: local predisposing factors. Ann Surg 1984; 200: 479-84. 96 Armstrong CP, Taylor TV, Jeacock J, Lucas S. The biliary tract in patients with acute gallstone pancreatitis. Br.J Surg 1985; 72: 551-5. 97 Stone HH, Fabian TC, Dunlop WE. Gallstone pancreatitis. Ann Surg 1981; 194:305-10. 98 Cuschieri A, CummingJGR, Wood RAB, Baker PR. Evidence ofsphincter dysfunction in http://gut.bmj.com/ patients with gallstone associated pancreatitis: effect of ceruletide in patients undergoing cholecystectomy for and gallstone associated pancreatitis. Br J Surg 1984; 71:885-8. 99 Toouli J, Roberts-Thomson IC, Dent J, Lee J. Sphincter of Oddi motility disorders in patients with idiopathic recurrent pancreatitis. Br J Surg 1985; 72: 859-63. 100 Mackie CR, Wood RAB, Preece PE, Cuschieri A. Surgical pathology at early elective operation for suspected acute gallstone pancreatitis: preliminary report of a prospective clinical trial. BrJ Surg 1985; 72: 179-81. on September 26, 2021 by guest. Protected copyright. 101 Armstrong CP, Taylor TV. Pancreatic-duct reflux and acute gallstone pancreatitis. Ann Surg 1986; 204: 59-64. 102 Cuschieri A, Howell Hughes J. Pancreatic reflux during operative choledochography. BrJ Surg 1973; 60: 933-6. 103 Cuschieri A, Howell Hughes J, Cohen M. Biliary-pressure studies during cholecystectomy. BrJ Surg 1972; 59: 267-73. 104 Mayer AD, McMahon MJ. Biochemical identification of patients with associated with acute pancreatitis on the day of admission to hospital. Ann Surg 1985; 201: 68-75. 105 Acosta JM, Pellegrini CA, Skinner DB. Etiology and pathogenesis of acute biliary pancreatitis. Surgery 1980; 88: 118-25. 106 Kelly TR. Gallstone pancreatitis: pathophysiology. Surgery 1976; 80: 488-92. 107 Acosta JM, Ledesma CL. Gallstone migration as a cause of acute pancreatitis. N Engl J Med 1974; 290: 484-7. % 108 Acosta JM, Rossi R, Galli OMR, et al. Early surgery for acute gallstone pancreatitis: evaluation of a systemic approach. Surgery 1978; 83: 885-8. 109 Farinon AM, Ricci GL, Sianesi M, Zanella E. Physiopathologic role of microlithiasis in gallstone pancreatitis. Surg Gynecol Obstet 1987; 164: 252-6. 110 Neoptolemos JP, Davidson BR, Valiance D, Winder AF. Bile crystal analysis for the detection of gall stones associated with acute pancreatitis. Br J Surg 1987; 74: 1160.