Abstract
The cellular mechanisms involved once pancreatitis has been initiated are reasonably well understood. The events leading up to this process are less well established. Much of our current understanding of pancreatitis in cats has been determined from experiments in cats or extrapolated from other species. The normal anatomy and function of the pancreas and a review of the current state of knowledge about the pathophysiology of pancreatitis is discussed. The current prevalence of feline pancreatitis is unknown, but the disease is being reported with increasing frequency. The aetiology of pancreatitis and the types of pancreatic inflammation present in cats is different from other species, such as the dog, a species where the disease is considered more common. Concurrent diseases are often present that may be more serious than the pancreatic inflammation and the treatment of these diseases is often complicated by pancreatitis.
Anatomy and physiology
Cats have different embryological development and anatomy of the pancreas from other species, including dogs. The pancreatic duct is derived from the ventral anlage and is the main functional pancreatic duct in cats, but it is of minor importance, and may be absent in dogs (Jubb 1993). In cats the accessory pancreatic duct generally does not persist, with 80% of cats having only one pancreatic duct (Jubb 1993). The accessory pancreatic duct enters the duodenum through the minor duodenal papilla. The pancreatic duct enters through the major duodenal papilla. In cats this is the main, and frequently the only, pancreatic duct opening into the duodenum, contiguously with the bile duct (Garvey & Zawie 1984, Jubb 1993, Williams 1996).
The primary role of the exocrine pancreas is to aid the digestion and assimilation of food and protect against auto-digestion. Pancreatic secretions include digestive enzymes to help break down lipids, proteins and polysaccharides in the proximal duodenum. Bicarbonate is secreted to neutralise gastric acid when food boluses enter the duodenum from the stomach. Colipase is secreted to facilitate the action of lipase in breakdown of fats. Factors are also secreted that enable the absorption of zinc and cobalamin (vitamin B12). Pancreatic secretions inhibit bacterial proliferation in the proximal small intestine and promote the normal degradation of exposed brush border enzymes (Williams 1996).
Protection against auto-digestion is complicated and there are many safeguards in place, our understanding of which is based primarily on human and laboratory animal physiology (Geokas et al 1985, Schaer 1991, Simpson 1993, Stewart 1994, Williams 1996). They are as follows:
The synthesis, transport and secretion of pancreatic enzymes into the duodenum occurs in an inactive form (zymogens). Once they enter the duodenum they are activated by enterokinase, an enzyme produced in the duodenal brush border.
Within the pancreatic cells the lysosomal enzymes (which may activate inert zymogens) are separated from zymogens by intracytoplasmic membranes. Immediately after synthesis of the pro-enzymes, the lysosomal enzymes are segregated from the cytoplasm by interposition of the endoplasmic reticulum, Golgi complex, zymogen granule membrane and the cell membrane itself.
A specific trypsin inhibitor of low molecular weight, pancreatic trypsin secretory inhibitor (PTSI) is present in pancreatic juice. It can inactivate free trypsin that may have been prematurely activated in the acinar cells or ducts.
Antiproteases such as alpha-antitrypsin, alpha-macroglobulin and anti-chymotrypsin circulate in the plasma and protect against proteases that escape into the circulation.
As such, activation of zymogens within the duodenum instead of within the pancreas is one of the major mechanisms of auto-protection by the pancreas (Rinderknecht 1986). Three steps have been described in this process (Rinderknecht 1986). The first two steps involve removal of the initiator methionine and then the transport peptide from the zymogen, these processes taking place in the rough endoplasmic reticulum. The third step involves cleaving of the activation peptide (inhibitor peptide) by enterokinase within the duodenal lumen. The brush borders of the proximal small intestine secrete enterokinase and it is 2000 times more effective at activating trypsinogen than trypsin itself (Rinderknecht 1986). Another biological characteristic of enterokinase is its enhanced activity in the presence of low bile acid concentrations and in environments with a pH range of 6–9 (Rinderknecht 1986). It is not inhibited by PTSI and does not form complexes with anti-proteases (Rinderknecht 1986).
Once trypsin has been activated by enterokinase, it in turn activates other pancreatic zymogens within the duodenal lumen. All of the zymogens (chymotrypsinogen, proelastase, kallikreinogen, procarboxypeptidase A and B, prophospholipase A and procolipase) are activated by trypsin, and they do not auto-activate (Rinderknecht 1986). All zymogens when activated cleave peptides, although sometimes they are not completely severed and can be attached by a disulfide bridge. The size of the activation (or cleavage) peptides varies (Rinderknecht 1986).
Although the exocrine pancreas secretes mainly in response to ingested food, there is still secretion into the duodenum during fasting (Williams 1996). This basal secretion accounts for approximately 2% of the bicarbonate and 10% of the digestive enzymes that are secreted in response to food (Williams 1996). There is a biphasic secretory response to food. The initial peak occurs 1–2 h after feeding and consists primarily of digestive enzymes. The second peak which occurs 8–11 h after feeding consists primarily of bicarbonate and is larger in volume (Williams 1994, Williams 1996).
Pancreatic secretion is mediated by a combination of neural and hormonal mechanisms. In cats the hormonal mechanisms are considered to be the most important (Williams 1996). Secretin and cholecystokinin (CCK) are released when food enters the small intestine and in turn stimulate pancreatic secretions. Secretin stimulates bicarbonate rich pancreatic juice production, while CCK causes enzyme-rich pancreatic juice production. Neural mechanisms occur via cephalic stimulation (anticipation and smell of food) and gastro-intestinal stimulation (presence and movement of food through the proximal gastrointestinal tract) (Williams 1996).
Although the majority of exocrine secretions reach the duodenal lumen via the pancreatic duct after stimulation, a small quantity enters the blood stream. There is no known mechanism for this event, but it means that some animals without pancreatic disease will have a small amount of circulating zymogens (Geokas et al 1985). Conversely, only in animals with pancreatic inflammation should there be circulating active enzymes.
Pathophysiology
The development of pancreatitis is widely thought to be due to failure of the protective mechanisms of the pancreas, resulting in zymogen activation within the pancreas (Rinderknecht 1986, Schaer 1991, Simpson 1993, Williams 1996, Luthen et al 1998). The initiating events that overcome or alter these defence mechanisms are less well understood. Many researchers have confirmed that trypsin activation within the pancreas is the initiating event in the development of pancreatitis and that tryspin is capable of activating other pancreatic zymogens (Nakae et al 1995, Mithofer et al 1995, Halangk et al 1997, Hofbauer et al 1998b, Luthen et al 1998, Klonowski-Stumpe et al 1998).
There is debate amongst researchers about the cause of trypsinogen activation. In summary, much research has been carried out and the following is known about the possible mechanisms for trypsinogen activation:
There is experimental evidence that zymogen granules and lysosomal hydrolases co-localise in vacuoles that appear within the cytoplasm. This is particularly true when the experimental model used is infusion of supra-physiological doses of cerulein, a cholecystokinin analogue (Simpson 1993, Frick et al 1997, Halangk et al 1997, Hofbauer et al 1998b, Mithofer et al 1998).
These vacuoles have been observed in healthy rats with no resulting activation of trypsinogen (Willemer et al 1990).
It has been shown in vitro that the lysosomal enzyme Cathepsin B activates trypsinogen to trypsin at pH 3.6. This process is enhanced by cysteine (a cathepsin activator) and inhibited by iodoacetic acid (a cathepsin inhibitor) (Greenbaum & Hirshkowitz 1961).
Recently chronic hereditary pancreatitis has been identified as a significant cause of pancreatitis in humans (Paolini et al 1998). It has been shown that there is a mutation of the cationic trypsinogen gene with an Arg-His substitution at residue 117. This residue site is a trypsin sensitive site, so its loss would permit auto-activation of trypsinogen, resulting in pancreatitis (Whitcomb et al 1996, Ferec et al 1999).
Experiments have shown that hyper-stimulation-induced pancreatitis can be prevented by a specific Cathepsin B inhibitor, E64D (Saluja et al 1997). However experimental models applying Cathepsin B inhibitors have not reduced activation of trypsinogen, but serine protease inhibitors have done so (Halangk et al 1997, Klonowski-Stumpe et al 1998).
Trypsinogen auto-activation occurs in vitro at pH 5.0 (Figarella et al 1988).
The exact pathophysiology of spontaneous pancreatitis in cats at a cellular level has not been confirmed, but it is assumed to be similar to that of dogs and humans (Simpson 1993, Williams 1996). Although this is speculation, pancreatitis has been experimentally induced in cats using similar models to other species like rats, mice and dogs and resulted in similar pathological consequences (Kitchell et al 1986, De Giorgio et al 1993, Reber et al 1998). Activation of trypsinogen has also been demonstrated experimentally in cats (Karanjia et al 1993). However, because feline pancreatitis differs in many ways clinically from both canine and human pancreatitis this inference may not be true in the natural course of the disease in cats.
Once pancreatic proteases have been activated and overwhelm the normal protective mechanisms of the pancreas they enter the interstitium of the pancreas and the peritoneal cavity (Schaer 1991). The progression of mild pancreatitis to severe disease is thought by many researchers to be due to the interstitial activation of trypsinogen (Fernandez-del-Castillo et al 1994, Hartwig et al 1999). Activated pancreatic enzymes may damage pancreatic tissue, and phospholipase A2 is considered to be one of the main enzymes responsible for this damage (Mayer et al 1998).
Circulating proteases activate the complement, fibrinogen, coagulation and kinin cascades (Lasson & Ohlsson 1984a, Schaer 1991, Williams 1996). The role of cytokines and other circulating inflammatory mediators in severe pancreatitis is generating significant research in order to develop substances that block interleukin-1 and tumour necrosis factor-α activity (Kusske et al 1996, Saluja & Steer 1999). Some studies have shown that use of agents that antagonise inflammatory mediators results in a decrease in the severity of pancreatitis (Van Laethem et al 1995, Hofbauer et al 1998a). Some reports also show that platelet-activating factor (PAF) plays an important role in systemic complications, particularly in lung associated disease in humans (Sandovel et al 1996, Hofbauer et al 1998a). The clinical implications of these experiments are that PAF antagonists such as lexipafant may reduce the degree of systemic complications even if they are administered after the onset of pancreatic inflammation.
The degree of damage to the pancreas and activation of systemic cascades is regulated by circulating plasma protease inhibitors (Lasson & Ohlsson 1984b). There are two major circulating plasma protease inhibitors: α1-protease inhibitor (α1-antitrypsin) and α-macroglobulins (α1-M and α2-M). These have been shown to decrease markedly in dogs with experimental pancreatitis and in people with acute pancreatitis, but no studies have been performed in cats (Lasson & Ohlsson 1984a, Lasson & Ohlsson 1984b, Murtaugh & Jacobs 1985). The action of α1-protease inhibitor is temporary, and its role is to transport proteases to the α-macroglobulins, especially from extravascular spaces where the large α-macroglobulin molecule cannot readily diffuse (Lasson & Ohlsson 1984a). Another secondary role of α1-protease inhibitor is to inhibit neutrophil elastase in inflammation (Williams 1996). The main protease binding function during pancreatitis is performed by the α-macroglobulins (Lasson & Ohlsson 1984a). Once proteases are bound the complex is cleared by the monocyte-macrophage system (Williams 1996). If the circulating pool of α-macroglobulins are exhausted then rapid development of systemic complications such as disseminated intra-vascular coagulation (DIC) result.
Systemic and local complications are common in acute pancreatitis (Williams 1996). Potential complications that can occur early in the course of pancreatitis are diabetes mellitus, diabetes ketoacidosis, pancreatic abscess/pseudocyst formation, cardiac arrhythmias, abdominal distension, ileus, disseminated intravascular coagulation (DIC), septicaemia, bile duct obstruction, respiratory distress and renal failure (Schaer 1991, Stewart 1994). Late onset complications include the development of exocrine pancreatic insufficiency or chronic relapsing pancreatitis, both of which have been documented in cats (Schaer 1991, Steiner & Williams 1996, Steiner & Williams 2000).
Pancreatic pseudocysts are collections of fluid containing pancreatic enzymes and debris that accumulate in a non-epithelialised sac (Bradley 1993). Pancreatic pseudocysts are common in humans as a consequence of pancreatitis, but they have been reported less frequently in cats (Hines et al 1996, Van Enkevort et al 1999). Percutaneous aspiration of pseudocysts is the treatment of choice in humans providing the risk of peritonitis is low, and the cystic lesion is not too large (Geokas and others 1985). Pancreatic pseudocysts need to be differentiated from other pancreatic masses such as abscesses or neoplasms. A true pancreatic abscess is a collection of septic exudate, generally resulting from secondary infection of pancreatic tissue or pseudocysts (Bradley 1993). Pancreatic abscesses, neoplasia and pseudocysts can all result in necrosis of the pancreas and subsequent destruction of glandular tissue and significantly worsen the prognosis (Bradley 1993).
Classification
Pancreatitis is generally classified clinically as acute, recurrent acute or chronic based on longevity and severity of clinical signs (Simpson 1993). Acute pancreatitis tends to be sudden in onset, have severe clinical signs and is more often associated with systemic complications (Simpson 1993, Williams 1996). Recurrent acute pancreatitis is repeated bouts of severe pancreatitis, while chronic pancreatitis tends to be milder. Histologically, recurrent acute pancreatitis results in little or no permanent pathological change, while there are irreversible morphological changes in chronic pancreatitis (Williams 1996).
Acute pancreatic inflammation is histologically described as acute pancreatic necrosis (APN) in order to characterise the necrotising nature of the lesions (Jubb 1993). The pathological process of APN is confined to the interstitium and peripancreatic adipose tissue, while the ducts and centrilobular tissues are spared during early phases of the disease (Jubb 1993). Moderate to severe pancreatic acinar and periacinar fat necrosis with minimal to moderate inflammatory infiltrate of neutrophils and macrophages with minimal to mild interstitial fibrosis is present (Hill & Van Winkle 1993, Jubb 1993). Necrosis often includes the adjoining adipose and mesenchymal tissue. Some authors further subdivide APN into those cases with or without fibrosis, although this classification contradicts the previous definition of recurrent acute versus chronic (Hill & Van Winkle 1993). Previously, acute pancreatitis in dogs was also described as haemorrhagic pancreatitis based on gross appearance of the organ, but this has been discarded in order to describe the histopathological changes present (Jubb 1993).
Chronic pancreatitis is the mildest form of pancreatitis and is characterised histologically by interstitial oedema with mild inflammatory infiltrate (Schaer 1991). Until recently chronic interstitial pancreatitis was thought to be the only form found in the cat (Jubb 1993, Simpson 1993). Although it is by the far the most common type, there have been reports of several cats with acute necrotising pancreatitis (Hill & Van Winkle 1993, Simpson et al 1994, Williams 1996, Steiner & Williams 1997, Swift et al 2000). A distinct form of suppurative pancreatitis has also been described in cats (Hill & Van Winkle 1993, Jubb 1993, Williams 1996). The histological findings of this form of pancreatitis are moderate to severe pancreatic inflammation (predominantly neutrophil involvement) with minimal acinar or fat necrosis (Hill & Van Winkle 1993).
One report of 40 cats diagnosed with acute pancreatitis at post-mortem identified 32 with APN and eight with acute suppurative pancreatitis based on histological classification (Hill & Van Winkle 1993). The presence of necrosis is considered to be an indicator of severity in people, dogs and cats (Bradley 1993, Hill & Van Winkle 1993, Williams 1996, Swift et al 2000). Due to the confusion over terminology the terms interstitial, necrotising and suppurative pancreatitis are best based on histological descriptions of the pancreatic inflammation. Clinical classification of pancreatitis should be based on the severity and longevity of the animal's clinical signs, although interstitial pancreatitis is usually mild in severity and chronic in longevity (Simpson 1993).
Prevalence
Ante-mortem diagnosis of pancreatitis in cats is uncommon because of its low incidence and difficulties associated with making a definitive diagnosis (Schaer 1991, Schaer & Holloway 1991, Simpson et al 1994). Feline pancreatitis is often diagnosed only at necropsy (Duffell 1975, Owens et al 1975, Jubb 1993, Simpson et al 1994). In one retrospective review of feline necropsies, approximately 3.5% had evidence of exocrine pancreatic disease (Dill-Macky 1993). This review may have overestimated the percentage of cats with pancreatitis as pancreatic lesions such as pancreatic neoplasia, amyloidosis, cysts and exocrine pancreatic insufficiency were included. Another review of 8687 cat necropsies over a 13-year period reported 40 with acute pancreatitis, accounting for one in 800 of sick cats presented to the authors' hospital (Hill & Van Winkle 1993). Other reports have recorded the incidence of feline pancreatitis to range from 0.57 to 2.9% (Spinaci & Marcato 1993, Steiner & Williams 1997). It is generally accepted that the interstitial form of pancreatitis is much more common in cats than the acute suppurative or necrotising forms (Schaer 1991, Hill & Van Winkle 1993, Simpson 1993, Williams 1996). It is also probable that there is a higher incidence of the disease in North America than Europe because of the higher prevalence of concurrent diseases such as hepatic lipidosis in that continent.
Concurrent disease
In a review of 40 cats with acute pancreatitis diagnosed at post-mortem 14 (35%) had mild nephritis and one cat (2.5%) with suppurative pancreatitis also had severe cholangiohepatitis (Hill & Van Winkle 1993). The mild nephritis was thought by the authors to be an incidental finding and directly related to the older age of the cats. In a review of 54 cats with inflammatory liver disease, cats with lymphocytic portal hepatitis did not have a higher incidence of pancreatitis than cats without hepatic disease (Weiss et al 1996). Conversely, inflammatory bowel disease and pancreatitis was present in 83% and 50% respectively of the cats with cholangiohepatitis in this review (Weiss et al 1996). All three diseases were present in 39% of cats with cholangiohepatitis. All cats with pancreatitis and concurrent cholangiohepatitis in this study had mild, interstitial pancreatic inflammation while the concurrent inflammatory bowel disease (based on duodenal biopsy) was severe. The pancreatitis was thought to be due to inflammation or blockage of the distal portion of the common bile duct that resulted in ascending infection or reflux of pancreatic secretions up the pancreatic duct. In another recent study 16 out of 18 cats with pancreatic disease had concurrent intestinal or hepatic disease (Swift et al 2000). In fact, 10 of these cats had pancreatic, intestinal and hepatic disease concurrently.
In a study of 13 cats with hepatic lipidosis by Akol et al (1993), five (38%) were also found to have acute pancreatitis. The presence of acute pancreatitis worsened the prognosis of these cats due to cachexia and the presence of concurrent coagulation abnormalities. Hill & Van Winkle (1993) also found fatty changes in the livers of 59% of cats with acute pancreatitis. The fatty liver may be due to anorexia associated with pancreatitis or be a coincidental finding. Concurrent hepatic lipidosis and pancreatitis have been identified in other studies (Bruner et al 1997, Swift et al 2000). It has also been shown that the presence of concurrent pancreatitis increases the mortality in cats with diabetic ketoacidosis (Bruskiewicz et al 1997). Additionally, cats with concurrent diabetes mellitus and chronic pancreatitis are more difficult to manage, with normal blood glucose values rarely attained by insulin administration, dietary changes and/or oral hypoglycaemic therapy (Goossens et al 1998).
Aetiologies
Feline pancreatitis has been known to occur after trauma such as falling from a height or road traffic accidents (Schaer 1991). Feline infectious peritonitis (FIP) and Toxoplasma gondii infection have also been associated with pancreatitis in cats, although they are rare causes (Smart et al 1973, Weiss & Scott 1981, 1993, Simpson 1993). In one recent review of 100 cats with toxoplasmosis 38 of the 45 cats that had a post-mortem examination had pancreatic inflammation (Dubey & Carpenter 1993). Infection with herpesvirus and parvovirus has been implicated, but never proven, to cause pancreatitis in cats (Steiner & Williams 1997). Aberrant migration of the trematode Amphimerus pseudofelineus is also a known, although uncommon, cause of pancreatitis in the cat (Steiner & Williams 1997). Topical organophosphate toxicity has been reported as a potential cause of feline pancreatitis (Hill & Van Winkle 1993). There is no evidence that profound hyperlipidaemia or hypertriglyceridaemia is a cause of pancreatitis in cats (Williams 1996). Steiner and Williams (1997) observed that 90% of cats with pancreatitis never have a definitive cause identified. Potential, although speculative, causes include nutrition, toxins, pancreatic ischaemia and duodenal/biliary reflux. Pancreatic colonisation by bacteria via haematogenous spread, transmurally from the colon and by reflux from the gall bladder has been experimentally documented in cats (Widdison et al 1994).
