This study reports the clinical, clinicopathological and ultrasonographic findings from dogs with chronic pancreatitis (CP). Fourteen dogs with clinical signs consistent with CP and histological confirmation of the disease were evaluated. Abdominal ultrasound and clinical pathology results were recorded. Sensitivities of pancreatic enzymes for diagnosis of CP were calculated with two different cut-off values. The mean age of affected dogs was 9.1 years. Spaniels were the most common breed with CP, representing seven of the 14 dogs in this study. CP was histologically severe in nine cases. Most dogs showed chronic low-grade gastrointestinal signs and abdominal pain. Five dogs had exocrine pancreatic insufficiency and five dogs had diabetes mellitus. The sensitivity of elevated trypsin-like immunoreactivity for CP was 17 per cent. The sensitivities of canine pancreatic lipase immunoreactivity, lipase and amylase for CP were 44 to 67 per cent or 14 to 28 per cent depending on the cut-off value used. Cholesterol was elevated in 58 per cent of samples. Liver enzymes were often elevated. The pancreas appeared abnormal on 56 per cent of ultrasound examinations. Ten dogs had died by the end of the study period; only one case was due to CP.
Statistics from Altmetric.com
CHRONIC pancreatitis (CP) in dogs is poorly documented clinically. However, a recent study reported the prevalence of CP to be 34 per cent in postmortem examinations of dogs from first opinion practice (Watson and others 2007). Another postmortem examination study found lymphocytic pancreatic lesions in 52.5 per cent of a selected population of dogs (Newman and others 2006). If postmortem lesions translate to clinically significant disease, CP must be much more common than currently recognised. This may reflect difficulties in non-invasive diagnosis and non-specific clinical signs, mirroring human CP (Etemad and Whitcomb 2001, Watson 2004). Often, human CP is not recognised until a late stage when endocrine and/or exocrine insufficiency have developed, leading to diabetes mellitus (DM) and exocrine pancreatic insufficiency (EPI), respectively (Etemad and Whitcomb 2001). In an attempt to improve the diagnosis of canine CP, a prospective observational study was undertaken to identify the clinical, clinicopathological and diagnostic imaging findings in dogs with histologically confirmed CP.
Materials and methods
Criteria for enrolment
All dogs presenting to the Queen's Veterinary School Hospital (QVSH) between January 2002 and March 2006 were considered for inclusion in the study. Dogs were initially identified with clinical signs suggestive of CP as previously described (Steiner 2003, Watson 2003, 2004, 2005, Xenoulis and others 2008). Dogs were subsequently enrolled if CP was confirmed histologically on pancreatic tissue taken either at surgery or postmortem examination. Concurrent clinical conditions, if any, were recorded but did not result in exclusion. Dogs did not receive any medication apart from that detailed in the results. Procedures were only performed if clinically justified, with fully informed owner consent.
Blood samples were taken following a minimum of a 12-hour fast into EDTA, plain and heparinised tubes in all cases on one or more occasions for complete blood count (CBC), routine biochemistry, lipase, amylase, canine trypsin-like immunoreactivity (TLI) and canine pancreatic lipase immunoreactivity (cPLI). Routine biochemistry included the following parameters: blood urea, creatinine, glucose, total protein, albumin, globulin, sodium, potassium, chloride, calcium, phosphate, alkaline phosphatase (ALP), alanine aminotransferase (ALT), aspartate aminotransferase and creatine kinase. CBC was performed within two hours using a Cell-Dyn 3500R analyser (Abbott Diagnostics), and a concurrent blood smear was evaluated. Biochemistry was performed within two hours using a Beckman CX5 Synchron Clinical System analyser (Beckman Coulter). Serum TLI was measured using a validated canine radioimmunoassay at the Cambridge Specialist Laboratory Services, Sawston, Cambridge. Serum for the measurement of cPLI was stored and shipped frozen and measured using a validated canine ELISA at the Gastrointestinal Laboratory, Texas A&M University, College Station, Texas, USA.
All dogs underwent pancreatic ultrasonography on one or more occasions. All dogs were conscious but some were sedated with low doses of acepromazine (acepromazine maleate injection 2 mg/ml; Novartis Animal Health) and buprenorphine (Vetergesic; Alstoe Animal Health) or butorphanol (Torbugesic; Fort Dodge Animal Health) given intramuscularly. Transcutaneous ultrasound was performed by an experienced operator using a HDI 5000 SonoCT machine (Philips Medical) with an 8 MHz curvilinear transducer or ATL 800 machine (Philips Medical) with a 7.5 MHz sector transducer. One of the authors (MEH), a board-certified radiologist, reviewed all of the images.
Tissue samples were obtained and processed at either QVSH or the referring practice or private pathology laboratories. Original blocks were obtained from all dogs and re-cut sections were reviewed by one of the authors (PJW) with the help of a veterinary pathologist (AJR or TJS), all blinded to the clinical details of the case, but aware that CP was suspected in each dog. At all postmortem examinations, three samples were taken from the pancreas, including right and left limbs and body. In all but two cases, samples were fixed within two hours of death. Case 11 (a seven-year-old neutered male cocker spaniel in thin body condition) was kept at 4°C for 12 hours before sampling (this was considered acceptable, as one human study has found that good-quality pancreatic tissue can be obtained up to 12 hours postmortem at room temperature) (Siriwardana and others 2010). Case 13 had been frozen at the referring practice immediately after death and was defrosted, and the pancreas and liver removed a day later. Surgical biopsies comprised one or two samples taken from the edge of one pancreatic lobe. Samples were fixed in 10 per cent buffered formalin, embedded in paraffin, sectioned at 4 µm, mounted on glass slides and stained with haematoxylin and eosin for histological examination and with Masson's trichrome for evaluation of fibrosis. CP was diagnosed, as previously described (Etemad and Whitcomb 2001, Watson and others 2007), by the presence of fibrosis distorting pancreatic architecture, together with the loss of acinar tissue and an associated mononuclear (macrophages, lymphocytes and plasma cells) or mixed (neutrophils and mononuclear cells) inflammatory infiltrate. Cases with fibrosis and inflammation involving between 10 and 40 per cent of the sections examined were considered equivalent to grade 2 disease and cases with over 40 per cent of the sections involved were considered equivalent to grade 3 disease (Newman and others 2006).
Diagnosis of EPI, DM and hyperadrenocorticism
EPI was diagnosed as previously described (Westermarck and Wiberg 2003) on the basis of a combination of consistent clinical signs (steatorrhoea, weight loss and response to dietary enzyme supplementation) and clinical pathology (low serum TLI in the diagnostic interval for EPI [<2.5 ng/ml] on more than one occasion). DM was diagnosed on the basis of persistent hyperglycaemia and glucosuria, and in some cases ketonuria, as previously described (Nelson 2004). Hyperadrenocorticism was diagnosed with a positive adrenocorticotropic hormone (ACTH) stimulation test supported by compatible clinical and ultrasonographic findings as previously described (Herrtage 2004).
Assessment of body condition
Body condition was assessed by one author (PJW) semiobjectively on enrolment by assessing body fat cover over the ribs and lumbar spine using a previously validated body condition scoring (BCS) system (Laflamme 1997, Mawby and others 2004). Dogs were categorised broadly as 1 Cachexic, 2 or 3 Thin, 4 or 5 Ideal, 6 or 7 Overweight, or 8 or 9 Grossly obese.
Most cases had frequent repeat visits to QVSH. Cases were followed up by telephone if they had not had a recent re-visit.
Sensitivities of TLI, cPLI, amylase and lipase for the diagnosis of CP were calculated from all samples taken from all cases except cases 9 and 11 and the last visit of 14. These were removed from analysis because they had hyperadrenocorticism (case 9) and renal failure (cases 11 and last visit of 14). Renal failure is known to falsely elevate amylase, lipase (Polzin and others 1983) and TLI (Mansfield and Jones 2000) and also to elevate cPLI (Steiner and others 2001). Corticosteroids are known to increase lipase and TLI and reduce amylase (Parent 1982, Lucena and others 1999). For diagnosis of pancreatitis, the established laboratory cut-off for TLI of >35 ng/ml was used. Two sensitivity values were calculated for cPLI, using established laboratory reference ranges (Steiner and others 2003) of either 102 ng/ml (‘less stringent’ interpretation) or 200 ng/ml (‘stringent’ interpretation) as the cut-off values (grey area 102 to 200 ng/ml); two values were also calculated for amylase and lipase using either the top of the laboratory reference interval as the cut-off value (less stringent interpretation) or using three times the top of the reference interval as the cut-off value (stringent interpretation) as previously suggested (Strombeck and others 1981, Steiner 2003). The sensitivity of finding any abnormalities on ultrasound for the diagnosis of CP was also calculated for all cases where the pancreas was imaged. Either a normal pancreas or no pancreas found on ultrasound was classed as ‘normal’.
As there was no control group, specificity and positive and negative predictive values could not be calculated.
Signalment and presenting clinical signs
A total of 61 dogs, suspected of having CP with suggestive clinical signs, were initially recruited to the study. Clinical signs considered as suggestive were chronic low-grade typical gastrointestinal (GI) signs or posthepatic jaundice with imaging suggesting pancreatitis. GI signs were intermittent anorexia, eating grass, vomiting, cranial abdominal pain and mild diarrhoea (often with some haematochezia), recurring over months to years, typically once a month or less and usually mild but sometimes severe enough to require hospitalisation and intravenous fluids. Fourteen of these 61 dogs were subsequently confirmed with CP on pancreatic histology and were included in this study. The clinical details of the 14 histologically confirmed dogs are shown in Table 1, together with any concurrent diseases. The mean (sd) age was 9.1 (2.5) years, with a median of 9.0 years. There were seven females and seven males. Spaniels were the most common breed with CP, representing seven of the 14 dogs (five English cocker spaniels and two Cavalier King Charles spaniels).
Ten dogs had a long history of mild, intermittent GI signs, two had acute posthepatic jaundice and two had polyphagia and weight loss. Three had pain on abdominal palpation. Three dogs with a chronic GI history also had acute GI signs just before referral. Polydipsia-polyuria (PD/PU) was reported in one dog with untreated hyperadrenocorticism and in another five dogs that did not have DM or other clinical signs suggestive of hyperadrenocorticism. In three of these dogs, ACTH stimulation tests were performed before referral. The results were negative in one dog and positive in two dogs, but these were later considered to be false positives, attributed to stress in both dogs in the absence of other compatible clinical signs of hyperadrenocorticism in either dog by the end of the study period. Neither dog showed any response to medication for hyperadrenocorticism administered by the referring veterinary surgeon and there was no change in signs when the medication was stopped. The PD/PU resolved on subsequent dietary management of CP.
Five of the 14 dogs had EPI. Two of these dogs also had DM and were both thin. In addition, one dog had diabetic ketoacidosis (DKA) on admission but likely no EPI, although it was thin and did not survive long enough to allow diagnostic testing to rule out EPI. A fourth dog was overweight and had newly diagnosed DM on admission but not EPI, and a fifth dog developed DM two years after presentation. In summary, by the end of the study, five of the 14 dogs had DM.
Nine dogs had identified concurrent diseases (apart from biliary obstruction and DM or EPI): one had hyperadrenocorticism, one had bacterial endocarditis and inflammatory bowel disease (IBD), two had mitral valve disease (one of which also had epilepsy), one had prostatitis, one had peripheral neuropathy and three had keratoconjunctivitis sicca (KCS).
Details of the body condition of the dogs on enrolment are shown in Table 1. Seven dogs were thin. All but one of these had DM or EPI, or both. An eighth dog was thin on the second visit when a hepatocellular carcinoma was diagnosed. Four dogs were overweight: none of these had EPI but one had DM. Three dogs were in ideal body condition and had neither EPI nor DM.
Clinical pathology: pancreatic enzymes and cholesterol
The results for TLI, cPLI, amylase and lipase are shown in Table 2. Values of pancreatic enzymes fluctuated widely between dogs and in different blood samples taken on different occasions from the same dog. In dogs with EPI, it was notable that TLI went up into the normal range or even into the range diagnostic for pancreatitis on occasions. TLI was in the diagnostic range for EPI in dogs with EPI on nine of 11 occasions that it was measured. cPLI was measured concurrently with a low TLI on six occasions in these dogs and only once was it in the range previously reported in dogs with EPI (less than 2.2 ng/ml) (Steiner and others 2006). Once it was in the range diagnostic for pancreatitis, and in the others it was normal.
Fasting cholesterol was elevated on 15 of 26 occasions that it was measured (58 per cent). In seven cases, concurrent conditions that are known to elevate fasting cholesterol were seen (Duncan 2005): DM (four times), hyperadrenocorticism (twice) and suspected cholestasis (once). All dogs with EPI had normal cholesterol from the time EPI was diagnosed, except case 8, which had elevated cholesterol at the first visit.
Serum triglyceride was measured on seven occasions but was elevated only once in one dog (case 8), which was also on phenobarbitone and potassium bromide treatment for epilepsy. The triglyceride concentration was 16.6 mmol/l (reference range 0.7 to 1.3 mmol/l) and the lipaemia may have interfered with the TLI measurement causing a false elevation. A repeat blood sample taken five days later without lipaemia showed a low TLI consistent with EPI.
Other clinical pathology
ALP was above the reference interval on 18 of the 20 occasions it was measured. Mean ALP was 1190 (1254) iu/l (removing the two cases with known causes of ALP induction, case 8 [on phenobarbitone] and case 9 [with hyperadrenocorticism], from analysis). ALT was also mildly elevated in many cases. In most dogs, the biochemistry and CBC were unremarkable apart from raised hepatic enzymes and cholesterol. Four dogs had mild to moderate neutrophilia; three dogs had mild hypoalbuminaemia (cases 1, 5 and 8; all thin spaniels and all with EPI); two dogs had mildly elevated urea with normal creatinine on one occasion, typical of prerenal effects, but normal urea on other occasions; and two dogs (cases 11 and 14 on the second visit) had marked azotaemia.
All dogs underwent pancreatic ultrasonography. The pancreas could not be identified on ultrasound on two occasions: case 1 on the only occasion it was imaged, and case 5 at the last visit. Both dogs had EPI at the time of ultrasound examination. On another nine occasions in another six dogs, the pancreas appeared unremarkable on ultrasound. Two of these dogs had EPI (cases 6 and 8). The remaining dog with EPI (case 4) had an abnormal pancreas on ultrasound. On 14 occasions, in 11 dogs, the pancreas appeared ultrasonographically abnormal, with a variety of findings ranging from a mixed hyperechoic and hypoechoic pattern (Fig 1) to a more uniform hypoechoic pattern and mass lesions that could be consistent with neoplasia in three cases. Neoplasia was ultimately ruled out in all three dogs based on pathological results, supported by a prolonged survival time in two dogs (the third was lost to follow-up). Ultrasound abnormalities showed no obvious predisposition to either pancreatic limb or the body. Three cases had abnormalities of the intestine and mesentery adjacent to the pancreas, including thickening of the small intestine and/or stomach wall, loss of layering and an apparent ‘mass’ effect invading the duodenum.
Sensitivities of diagnostic tests
The sensitivities of elevated values of TLI, cPLI, amylase and lipase for the diagnosis of pancreatitis and the sensitivity of finding an abnormality on ultrasound for the diagnosis of CP are shown in Table 3. Finding an abnormality on ultrasound had a sensitivity of 56 per cent. Elevated values of TLI, cPLI, lipase and amylase had sensitivities of between 17 and 67 per cent for the less stringent interpretation and sensitivities of between 14 and 28 per cent for the more stringent interpretation for pancreatitis.
Gross and histological pathology
The findings on pancreatic pathology in all cases are summarised later on in this article and illustrated in Figs 2 and 3. Five cases had surgical pancreatic biopsies taken by the referring practice, two cases had surgical biopsies at the QVSH, five cases had pancreatic sections taken at postmortem examination at QVSH and two cases had postmortem examination at the referring practice.
A macroscopic description was provided for 13 cases. In three cases (cases 1, 5 and 8), the pancreas was not grossly visible at surgery or postmortem examination but was completely replaced by mesenteric fibrosis with multiple adhesions to other visceral organs. All three of these cases had end-stage disease on histopathology and all three had EPI. Two of the three cases also had DM. In four cases (7, 12, 13 and 14), the pancreas appeared grossly normal despite significant histological CP. None of these cases had EPI and only one had DM.
Histological changes in the pancreas were severe, with gross disruption and loss of acini in nine cases (equivalent to grade 3 lesions reported by Newman and others 2006), and mild to moderate in five cases (all equivalent to grade 2 lesions reported by Newman and others ). Microscopic changes were diffuse in all tissues examined, in all cases except for case 2 where they were confined to one limb. The distribution of fibrosis and inflammation differed between breeds, with Cavalier King Charles and cocker spaniels having a perilobular pattern, while other breeds had an intralobular distribution (Fig 3). Pancreatic histology of case 4 was consistent with pancreatic acinar atrophy (PAA) and CP. The pancreas appeared firm and grossly nodular, and histologically there was acinar atrophy, but also marked fibrosis, ductular hyperplasia and many intraepithelial lymphocytes. The dog had a low serum TLI but no clinical signs of EPI until recently. A section of the small intestine from this dog showed eosinophilic enteritis and gastric biopsy revealed helicobacter and lymphoplasmacytic gastritis.
All cases were treated with a low-fat diet as well as insulin and/or pancreatic enzyme supplementation as necessary. One case was not regularly medicated or fed a low-fat diet because of poor owner and dog compliance. One case received antibiotics for endocarditis, one case received long-term antiepileptic medication and one case received trilostane (Vetoryl; Dechra Veterinary Products) for hyperadrenocorticism.
Outcome of cases
The outcome of cases is shown in Table 4. By the end of the study period, 10 dogs were dead, two were alive and doing well, one was lost to follow-up and one was doing well three years after enrolment when the dog ran away from home. Only one dog died as a direct result of pancreatic disease (case 11). In one case, the cause of death was unclear. In the other seven dogs, the cause of death was apparently unrelated to the pancreas. Four dogs died or were euthanased because of hepatic disease or neoplasia.
This is the first study to describe clinical, clinicopathological and ultrasonographic findings in a group of dogs with histologically confirmed CP. The pathology study reported by the authors suggests that CP is common (Watson and others 2007). This study provides evidence that CP is also clinically important in dogs. The most obvious clinical consequences were development of exocrine and/or endocrine insufficiency in some cases.
All of the dogs were middle-aged to old at presentation, mirroring the findings in human beings (Etemad and Whitcomb 2001).
No calculation of disease incidence or breed distribution was attempted in this study because the population of dogs was very biased. No standards for non-invasive diagnosis of CP in dogs have been published so the inclusion criterion for this study was very strict, requiring histological confirmation of disease, which meant that many potential cases of CP were excluded. At least some of the other 61 cases initially recruited were also likely to have had CP. This makes up a relatively large number of cases recruited over a four-year period and provides support for finding a high prevalence in pathology studies (Newman and others 2004, Watson and others 2007). There were a large number of spaniels in the current study, mirroring a pathology study (Watson and others 2007) that found an increased relative risk of pancreatitis in Cavalier King Charles and cocker spaniels, providing further evidence for breed-related disease in spaniels. Two studies have demonstrated an increased risk of EPI in older Cavalier King Charles spaniels in the UK (Hall and others 1991, Batchelor and others 2007), suggesting that this may be due to CP, although no pancreatic histopathological analysis was performed in those studies. The current study demonstrated histologically confirmed CP in Cavalier King Charles spaniels with EPI and also DM, providing evidence that CP is at least one cause of EPI in this breed.
The presenting clinical signs were reported carefully because these are first to alert the clinician to possible CP. Most dogs had mild, intermittent GI signs, which might be overlooked or confused with mild IBD or other low-grade GI disease (Jergens 1999, German and others 2003). However, the signalment should increase suspicion for CP rather than IBD. These dogs were typically middle-aged, small to medium breeds, whereas IBD typically affects young adult, large breeds such as German shepherd dogs (Jergens 1999, German and others 2003). However, the two diseases are not mutually exclusive, and some cases may have had both IBD and pancreatitis (such as case 4 reported here) as has been reported in human beings where some forms of CP occur concurrently with IBD (Zamboni and others 2004). Most of these dogs did not have small intestinal biopsies so concurrent IBD could not be ruled out. However, most dogs also responded clinically to a change to a low-fat diet without any anti-inflammatory, antibiotic or hypoallergenic treatments, suggesting that IBD was not an important clinical component of their condition because IBD, by definition, should not respond totally to dietary or antibiotic treatment without additional anti-inflammatory medication (Washabau and others 2010).
Abdominal pain was found on palpation in some of these dogs, and also sometimes noted by the owners. CP may be a significant and underestimated cause of morbidity in dogs. Increased recognition is important to allow appropriate management with diet and analgesia. Chronic epigastric pain is common in CP in human beings (Etemad and Whitcomb 2001, Hawes 2002). It does not necessarily correlate with histological severity of the disease (Walsh and others 1992) and is sometimes so severe as to lead to surgery to remove the pancreas (Behrman and Mulloy 2006). Postprandial pain appears to improve after feeding a low-fat diet to some human patients (Sibert 1978) and this also appeared to be the case in some of these dogs.
Nine dogs had concurrent disease at the time of presentation (in addition to EPI, DM and biliary obstruction, which were probably directly related to CP). Such a high prevalence of concurrent disease is not unusual in an old dog population and mirrors another study of canine pancreatitis published in 2003 (Mansfield and others 2003) in which 50 per cent of dogs had concurrent disease. Most concurrent diseases, such as KCS, mitral valve disease and peripheral neuropathy, have very different clinical signs from CP. However, one dog had concurrent prostatitis and one had concurrent IBD, which may have produced clinical signs similar to CP.
Three cases presented with ‘acute’ disease clinically, but the pancreas was already severely reduced in size and function, including one dog with no previous history of pancreatitis before presentation in terminal DKA crisis, which had virtually no normal pancreas left at postmortem examination. In human beings, individuals with CP often already have extensive loss of pancreatic tissue before presenting with their first clinically acute bout of disease (Renner and others 1985, Steer and others 1995). The present study confirms that this also occurs in dogs.
An association between obesity and acute pancreatitis has previously been reported in dogs (Cook and others 1993, Hess and others 1998, 1999), but the authors' pathology study found no significant association between obesity and CP, although 50 per cent of dogs with CP were overweight at postmortem examination (Watson and others 2007). The current study does not provide any convincing evidence either for or against an association between obesity and CP.
PD/PU in the absence of DM or hyperadrenocorticism was a noticeable clinical finding in four dogs. The cause of the PD/PU in these cases is unclear as none of them developed DM by the end of the study and the PD/PU resolved after a diet change. PD/PU of unknown pathophysiology has also been reported in some cases ofcanine gastrointestinal disease (Henderson and Elwood 2003). The clinical importance is that stress associated with ongoing pancreatic inflammation can lead to false-positive ACTH stimulation test results and, as in two cases, even inappropriate treatment, particularly as dogs with CP may also have polyphagia due to EPI, further mimicking hyperadrenocorticism. It is therefore important for clinicians to be aware of this possibility.
The time period between measurements of serum pancreatic enzymes and histological diagnosis was variable in these cases (as detailed in Table 2). Ideally, blood tests and biopsies should have been taken at the same time. However, this was not possible in all cases. Three of five cases that had a long time period between blood samples and histology had exocrine and/or endocrine insufficiency by the time of the first visit, suggesting there was already considerable loss of pancreatic mass. CP is a chronic disease resulting in permanent histological changes so it is likely that all cases had CP at the time of sampling. However, because the time course of CP in dogs is unknown, it is impossible to be sure of this.
Pancreas-specific immunoassays such as TLI and cPLI are considered superior to non-organ (or species)-specific catalytic enzyme assays such as amylase and lipase in the diagnosis of both canine EPI and canine pancreatitis (Steiner 2003, Watson 2005). However, in these 14 dogs, the immunoassays were no more sensitive than the catalytic assays and none of the available diagnostic tests (including ultrasound) had a high sensitivity for diagnosis of CP. The sensitivities of TLI, lipase, amylase and cPLI for the diagnosis of pancreatitis were very similar for the stringent tests, indicating no apparent benefit of measuring cPLI instead of lipase. However, cPLI may be more specific than lipase because it is pancreas specific (Steiner and others 2003), but this hypothesis could not be tested without a control group.
TLI was only low on nine of 11 occasions that it was measured in the dogs with EPI in this study. This is likely to be due to its concurrent elevation in some cases by ongoing pancreatic inflammation. This underlines the difficulty of using a single TLI measurement to diagnose EPI in dogs with CP, in contrast to using it in dogs with PAA and no associated inflammation, where a low value has a reported sensitivity of 100 per cent (Williams and Batt 1988).
Only one of the dogs in this study with EPI had a low cPLI on one occasion, in contrast to a study by Steiner and others (2006), which reported that all 25 dogs with EPI had a cPLI below the lower limit of the reference interval. In the current study, cPLI was often elevated in dogs with EPI. The marked difference in results between the two studies could be explained by the difference in the aetiology of the EPI: in the previous study (Steiner and others 2006), most of the dogs (18 of 25) were German shepherd dogs (identified in an earlier related paper (Rutz and others 2004) and were therefore likely to have non-inflammatory PAA. In the present study, all of the dogs with EPI had CP.
Fasting serum cholesterol and triglyceride concentrations were measured to assess the much-debated associations between lipaemia and hypercholesterolaemia and pancreatitis in dogs (Whitney and others 1987, Hess and others 1998). Cholesterol was elevated on many occasions, whereas triglyceride was only elevated in one of seven samples. In previous studies of naturally occurring acute pancreatitis (Schaer 1979, Hess and others 1998), 48 to 80 per cent of dogs had elevated cholesterol and 26 per cent had lipaemic serum. In this study, in some cases, the increase in cholesterol was probably due to concurrent, sometimes related, conditions (DM, hyperadrenocorticism and cholestasis). In other cases, no concurrent conditions were identified.
The finding of normal triglyceride levels on six of seven occasions was unexpected in the light of previous reports. However, this could be a particular feature of dogs with CP where concurrent loss of exocrine function results in fat maldigestion. This would be expected to affect serum triglycerides, which partly originate from chylomicrons absorbed from the GI tract, much more than cholesterol, which is also produced endogenously (Duncan 2005).
Other clinicopathological findings in this study were similar to those reported in studies of acute pancreatitis in dogs (Schaer 1979, Hess and others 1998). The most consistent findings were elevations in ALP and mild to moderate elevations in ALT, which have been previously reported to be elevated in 79 and 61 per cent of cases of acute pancreatitis, respectively (Hess and others 1998). In contrast, neutrophilia and azotaemia were uncommon in this study but are reported to occur in 55 to 60 per cent and 53 to 65 per cent of acute cases, respectively (Schaer 1979, Hess and others 1998). This difference probably reflects differences in disease severity between studies.
No previous studies report the results of transcutaneous ultrasound of a number of dogs with CP. Most report naturally occurring or experimentally induced acute pancreatitis where pancreatic ultrasound has a high specificity but lower sensitivity for diagnosis of inflammation, with typically a hypoechoic pancreas surrounded by hyperechoic mesentery (Saunders 1991, Simpson and others 1995, Hess and others 1998). It is likely that low-grade CP with predominant fibrosis and very little active inflammation produces very little ultrasonographically visible change and hence the low sensitivity of this method for diagnosis. Both dogs in which the pancreas could not be found had EPI. This was probably due to the marked reduction in pancreatic mass in these dogs with end-stage disease. The low sensitivity of transcutaneous ultrasound may be contributing to the under-recognition of CP. The ultrasonographers in this study were not blinded and may have been biased by clinical assumptions. It is therefore all the more significant that so many dogs had no ultrasonographic pancreatic lesions.
Three dogs in this study showed ultrasonographic lesions in their gut wall. In some cases, the loss of layering (presumably due to oedema and/or adhesions) and surrounding inflammation gave a mass effect, which could have been mistaken for neoplasia. In three cases, the appearance of the pancreas was also suggestive of neoplasia. In human beings, it is recognised that CP may mimic pancreatic neoplasia on diagnostic imaging, which may result in unnecessary removal of the pancreas at surgery (Etemad and Whitcomb 2001, Zamboni and others 2004). The mass-like appearance is due to a combination of inflammatory and fibrous tissue (Bartholomew and others 2006). This phenomenon has not previously been specifically reported in dogs, although mass-like lesions caused by CP resulting in extrahepatic biliary obstruction have been reported at surgery (Matthiesen and Rosin 1986).
In three dogs, the pancreas was not grossly visible at surgery or postmortem examination and all three of these dogs had EPI with end-stage disease on histology and very few acini remaining. In contrast, none of the four dogs with a normal gross appearance of the pancreas had EPI. It therefore appears that, as expected, pancreatic tissue was lost in parallel with the loss of exocrine function. The pancreas has a large exocrine and endocrine reserve and needs to lose 80 to 90 per cent of lipase production and islet cell function to develop EPI and DM, respectively (DiMagno and others 1973, Larsen 1993). Two of the three dogs with no visible pancreas had DM, but one did not and one of the dogs with a normal pancreas grossly had DM. This suggests that the relationship between loss of total pancreatic mass and loss of islet cell function is less straightforward. It may be that in some dogs, islets are relatively spared as acini are lost and that this depends on the aetiology of the CP, but this requires further investigation.
It is possible that CP is a common contributor to the aetiology of DM and EPI in dogs, but this is often not recognised because of the mild, non-specific clinical signs, the low sensitivity of available non-invasive tests and the lack of pancreatic histology in most cases (Watson 2003, Watson and Herrtage 2004). It is important to increase clinician awareness of end-stage CP because the lack of recognition of EPI in older individuals of breeds other than the German Shepherd dog with PAA can lead to a lack of diagnosis and treatment, which can result in unnecessary cachexia and euthanasia, particularly in dogs with concurrent DM as reported by Watson (2003).
It is also important to note that the pancreas can appear grossly normal and yet have histologically and clinically significant CP lesions. This has previously been reported in a pathology study (Newman and others 2004). This therefore emphasises the importance of taking biopsies in clinical cases with suggestive signs, even if the pancreas appears grossly normal.
The distinctive histological features of CP in dogs have been previously described (Newman and others 2004). There are some important differences from the histological appearance of PAA, which allow differentiation of the two diseases: PAA results in acinar atrophy and early lymphocytic infiltrate but no fibrosis or islet loss, whereas CP involves predominant fibrosis and inflammation affecting acinar, islet and interstitial tissue. It was therefore possible to define case 4 as having concurrent PAA and CP. There were also breed differences in the histological distribution of fibrosis and inflammation in the present study, which mirrored those reported in the authors' pathology study (Watson and others 2007) and suggests a different aetiology in different breeds, with spaniels showing a perilobular pattern and the labrador and small terrier-type breeds showing an intralobular pattern.
The outcome was good in these dogs and most animals did not die as a direct result of their disease. Even dogs with DM and EPI had good long-term outcomes with appropriate therapy. However, intriguingly, a surprisingly large number of dogs died or were euthanased because of liver disease, either liver failure or hepatic tumours. This high prevalence is unusual because primary liver tumours are rare in dogs, representing less than 1.5 per cent of all canine neoplasms (Liptak 2007), although, of all primary liver tumours, hepatocellular carcinomas are the commonest, comprising about 50 per cent (Liptak 2007).
In conclusion, this study demonstrates that CP is clinically significant in dogs and shows that some apparently ‘acute’ pancreatitic cases are in fact acute exacerbations of previously unrecognised chronic disease. CP has an important role to play in the development of both DM and EPI in dogs. It also appears to have an increased incidence and distinctive histological appearance in certain breeds, particularly spaniels, but no single diagnostic test that is currently readily available has a high sensitivity for the diagnosis of canine CP.
The results of some of these cases were presented at the ECVIM Congress in Amsterdam in September 2006 and form chapter 6 in PJW's VetMD thesis, University of Cambridge.
Clinical pathology was partly funded by a grant from the PetPlan Charitable Trust. Some of the pathological studies were funded by a grant from BSAVA Petsavers. The authors thank the dog owners and referring veterinarians, particularly Rob Foale, Tina Davy, Rob Davies and Victoria Roberts. They also acknowledge Madelaine Fordham and Rayner Speed for all their invaluable help processing the pathology samples. A number of external pathology laboratories in the UK were very helpful in providing blocks. The authors also acknowledge the help of Andrew Holloway, Lizza Baines and Nic Haywood in the ultrasonography of these cases.
Provenance not commissioned; externally peer reviewed
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.