Although pancreatitis is a relatively common disorder in dogs, the antemortem diagnosis of this disease remains clinically challenging. Clinical signs are protean and nonspecific, and can include anorexia, vomiting, lethargy, diarrhea, melena, weight loss, hematemesis, and hematochezia. Risk factors for acute pancreatitis in dogs include obesity, diabetes mellitus, hyperadrenocortcisim, hypothyroidism, prior gastrointestinal disease, and epilepsy, and Miniature Schnauzers and Miniature Poodles appear overrepresented.
Many diagnostic procedures and tests have been evaluated for their sensitivity and specificity for pancreatic disease detection. Measurement of serum lipase and amylase activity has been shown to lack specificity following the demonstration that pancreatectomy does not significantly alter their activity. In addition, increases in serum lipase activity occur in dogs with renal and hepatic disease in the absence of pancreatic inflammation.[4, 5] Measurement of serum trypsin and trypsinogen-like immunoreactivity (TLI) is highly sensitive and specific for the diagnosis of exocrine pancreatic insufficiency; however, this assay has limited value for the diagnosis of canine pancreatitis.[6, 7] More recently, a radioimmunoassay (RIA) and enzyme-linked immunosorbent assay (ELISA) were developed and validated for the measurement of canine pancreatic lipase (cPL).[8, 9] The canine pancreatic lipase immunoreactivity (cPLI) assay has immunoreactivity specific for pancreatic acinar cells. The cPLI assay has been refined by using monoclonal antibodies in a sandwich ELISA and recombinant antigen for calibration, and is now available commercially as the Spec cPL (canine pancreas-specific lipase) assay. The Spec cPL assay has a sensitivity of 63.6% in necropsied dogs with macroscopic evidence of pancreatitis; however, the specificity was not reported in that study. To the authors’ knowledge, there is only 1 recently published study reporting the specificity of the Spec cPL assay for diagnosing pancreatitis in 40 dogs without clinical or histopathologic evidence of the disease. The reported specificity of the Spec cPL in that study was 97.5% (95% confidence interval [CI], 86.8–99.9%).
The study objectives were to determine the sensitivity and specificity of several serum markers including Spec cPL, lipase, amylase, and cTLI for the diagnosis of pancreatitis in dogs using pancreatic histopathology on freshly procured pancreata from necropsied dogs as the “gold standard” of diagnosis.
- Top of page
- Materials and Methods
Dog breeds were diverse and included 10 mixed breed dogs, with no apparent breed predilections noted. Thirty-four of the dogs included in this study were male (3 intact, 31 neutered) and 36 were female (5 intact, 31 spayed). The dogs’ ages ranged from 9 months to 15 years (mean 8.5 years, median 9.2 years).
For analysis of pancreatitis, the best composite statistic used was the cumulative score (mean) and this metric was utilized for all of the statistical comparisons. The data set was analyzed in several other ways by calculating the average of the mean in which the mean score was calculated for the left limb, body, and right limb; the average of the maximum score in which the highest score identified for all transverse sections in each limb was calculated for the left limb, body, and right limb; and the maximum score for each parameter for the entire pancreas. Results of these composite scoring systems were similar to those derived from the cumulative score; hence the cumulative score was used for all comparisons.
Histopathologic evidence of mild pancreatitis was observed in 56 of 70 dogs, moderate pancreatitis in 6 of 70 dogs, and severe pancreatitis in 1 of 70 dogs. Seven of 70 dogs had no histopathologic evidence of pancreatitis. Virtually all the 63 dogs with pancreatitis had histopathologic features of “acute” and “chronic” pancreatitis concurrently, and only 5 dogs had histopathologic features of acute pancreatitis alone. For this reason, the statistician and coauthor on the article (PHK) concurred that the groups be combined to increase the power of our study. For the 7 dogs with no histopathologic evidence of pancreatitis, 6 had a Spec cPL < 200 μg/L and 1 had a Spec cPL between 200 and 400 μg/L. For the 56 dogs with histopathologic evidence of mild pancreatitis, 32 had a Spec cPL < 200 μg/L, 12 had a Spec cPL between 200 and 400 μg/L, and 12 had a Spec cPL > 400 μg/L. For the 7 dogs with histopathologic evidence of moderate to severe pancreatitis, 2 dogs had a Spec cPL < 200 μg/L, and 5 dogs had a Spec cPL > 400 μg/L.
Eleven of the 17 dogs (65%) with Spec cPL concentrations > 400 μg/L had clinical signs supportive of pancreatitis (vomiting, anorexia, weight loss), and 6 of the 17 dogs (35%) had physical examination findings of abdominal pain. No clinical signs or physical examination findings were reported in the medical records for 2 dogs. Six of these 17 dogs had macroscopic evidence of pancreatitis, with 4 of 6 having histopathologic evidence of moderate to severe pancreatitis. Additional pathological findings for 16 of 17 dogs included disseminated neoplastic disease (9), hepatic cirrhosis (2), immune mediated hemolytic anemia (1), glomerulonephritis (1), disseminated cryptococcosis (1), pneumonia (1), and myelodysplasia (1).
Serum amylase was the least sensitive serological assay for the diagnosis of pancreatitis with a sensitivity of 7% (95% CI = 2–17%) for mild pancreatitis, and 14% (95% CI = 0.4–58%) for moderate to severe pancreatitis, whereas serum lipase was the least specific serological assay (42.9% [95% CI = 10–82%]) for dogs with mild and severe pancreatitis (Table 1). The Spec cPL and serum lipase assays were the most sensitive assays for the diagnosis of moderate to severe pancreatitis (71% [95% CI = 29–96%]); however, the Spec cPL was more specific than all the other assays (100% [95% CI = 59–100%] and 86% [95% CI = 42–100%]) using a cutoff of 400 and 200 μg/L, respectively.
Table 1. Performance characteristics for the Spec cPL, cTLI, lipase and amylase in 70 dogs with and without histopathologic evidence of pancreatitis.
|Test||Outcome||Sensitivity (%) (95% CI)||Specificity (%) (95% CI)|
|(Cutoff >400 μg/L)||Mild pancreatitis||21 (12–34)||100 (59–100)|
|Moderate to severe pancreatitis||71 (29–96)||100 (59–100)|
|(Cutoff >200 μg/L)||Mild pancreatitis||43 (30–57)||86 (42–100)|
|Moderate to severe pancreatitis||71 (29–96)||86 (42–100)|
|cTLI (>35 μg/L)||Mild pancreatitis||30 (19–44)||100 (59–100)|
|Moderate to severe pancreatitis||29 (4–71)||100 (59–100)|
|Amylase (>1240 U/L)||Mild pancreatitis||7 (2–17)||100 (59–100)|
|Moderate to severe pancreatitis||14 (0.4–58)||100 (59–100)|
|Lipase (>750 U/L)||Mild pancreatitis||54 (40–67)||43 (10–82)|
|Moderate to severe pancreatitis||71 (29–96)||43 (10–82)|
The correlation coefficients for the Spec cPL assay was highest for suppurative inflammation (0.45), peripancreatic fat necrosis (0.43), acinar cell necrosis (0.42), and edema (0.35); histopathologic features most consistent with acute pancreatitis (Table 2). Histopathologic features consistent with chronic pancreatitis such as fibrosis (0.34), atrophy (0.24), and lymphocytic inflammation (0.16) had a poorer correlation with the Spec cPL. Despite these differences, the correlation coefficients for each of the 9 histopathologic features were highest for the Spec cPL assay compared to cTLI, serum amylase, and lipase (Table 2).
Table 2. Correlation coefficients for diagnostic tests and pancreatic histopathology using cumulative score in 70 dogs with and without histopathologic evidence of pancreatitis.
| ||Suppurative Inflammation||Acinar cell Necrosis||Hemorrhage||Interstitial Edema||Peripancreatic Fat Necrosis||Lymphocytic Inflammation||Fibrosis||Atrophy||Cellular Degeneration||Pancreatitis|
Twelve of 70 dogs had macroscopic evidence of pancreatitis at necropsy, including hemorrhage (8), edema (5), pancreatic nodules/plaques (6), a firm pancreas (4), hyperemia (2), pancreatic necrosis (1), and peripancreatic fat necrosis (3). The 12 dogs with macroscopic evidence of pancreatitis had significantly higher histopathologic scores for pancreatitis (P < .0001) compared to the 58 dogs without macroscopic evidence of pancreatitis. In addition, the Spec cPL was significantly higher in these 12 dogs compared to the other 58 dogs without macroscopic evidence of pancreatitis (P = .014). Five of 12 dogs with macroscopic evidence of pancreatitis had a histopathologic diagnosis of moderate to severe pancreatitis, whereas 7 of the 12 dogs had a histopathologic diagnosis of mild pancreatitis. Spec cPL was < 200 μg/L in 4 of 12 dogs, between 200 and 400 μg/L in 2 of 12 dogs, and > 400 μg/L in 6 of 12 dogs. Serum lipase activity was also significantly higher in the 12 dogs with macroscopic evidence of pancreatitis (P = .0005), although only 4 of these 12 dogs had abnormally increased serum lipase activity. In contrast, serum amylase activity and cTLI were not significantly different between the 2 groups of dogs (P = .17 and P = .12, respectively), and 4 of 12 and 5 of 12 dogs had abnormally increased serum amylase activity and cTLI concentrations, respectively.
Digital ultrasound images of the pancreas were available for interpretation in 8 of 28 dogs, and all of these dogs had ultrasonographic evidence of pancreatitis. In addition, all 8 dogs had histopathologic evidence of pancreatitis (3 dogs with mild pancreatitis and 5 dogs with moderate to severe pancreatitis). The Spec cPL was < 200 μg/L in 2 of these dogs, between 200 and 400 μg/L in 2 of these dogs and > 400 μg/L in 4 of these dogs. One dog with a histopathologic diagnosis of moderate pancreatitis had a Spec cPL of 76 μg/L, and 1 dog with a histopathologic diagnosis of mild pancreatitis had a Spec cPL of 77 μg/L.
Liver biopsies were examined in 12 dogs with Spec cPLs > 400 μg/L, and with histopathologic evidence of mild pancreatitis. Hepatic lesions in these dogs included: vacuolar hepatopathy (n = 9), lymphoma (n = 3), extramedullary hematopoiesis (n = 2), hepatocellular necrosis (n = 2), nodular hyperplasia (n = 2), hepatic congestion (n = 2), multifocal fibrosing portal lymphocytic hepatitis (n = 1), histiocytic sarcoma (n = 1), hepatocholangiocarcinoma (n = 1), metastatic islet cell carcinoma (n = 1), and hemorrhage (n = 1). Seven dogs had multiple lesions. The dog with metastatic islet cell carcinoma had pancreatic islet cell carcinoma and was found to have mild pancreatitis histopathologically. Intestinal biopsies were evaluated in 8 of these 12 dogs. Three of these dogs had abnormal intestinal pathology characterized by mural lymphoma of the duodenum (n = 1), mild segmental, plasmacytic enteritis (n = 1), severe necrosuppurative serositis (n = 1), mild focal acute hemorrhage of the ileum (n = 1), and crypt ectasia (n = 1). Two dogs had multiple lesions. Gastric biopsies were evaluated in seven of these 12 dogs, and abnormal findings were only evident in 1 dog that had evidence of severe submucosal edema of the stomach. Neoplasia in the pancreas was identified in 9 dogs: islet cell carcinoma (3), lymphoma (2), osteosarcoma (1), adenocarcinoma (1), mammary carcinoma (1), and transitional cell carcinoma (1).
- Top of page
- Materials and Methods
The Spec cPL was the most sensitive and specific test compared to all other serological markers assessed, with a sensitivity ranging between 43 and 71% for dogs with mild to severe pancreatitis, using a cutoff value of 200 μg/L. The sensitivity decreased for dogs with mild pancreatitis (21%) when the cutoff was increased to 400 μg/L; however, the specificity increased from 86 to 100% using a cutoff of 400 μg/L. The stringent histopathological definition of “normal pancreas” (ie, no histopathological changes in any examined section) resulted in the classification of relatively few dogs as being normal (7 of 70). The high specificity might be partially attributable to this relatively small population of “normal animals”. The sensitivity findings in dogs with moderate to severe pancreatitis are similar to results reported in a recent study evaluating the sensitivity of the cPLI in 22 dogs with macroscopic and histopathologic evidence of pancreatitis following necropsy; In that study, the cPLI had the highest sensitivity for dogs with macroscopic evidence of pancreatitis (64%) compared to other serological markers. The specificity of the Spec cPL has been published in 1 study to date, and the test was found to be highly specific (98%) in 40 dogs with no histopathologic evidence of pancreatitis. Our findings parallel those of Neilson-Carley et al, and the specificity of 100% (CI = 59–100%) using a cutoff of 400 μg/L suggests that false-positive test results are uncommon with the test.
The lower sensitivity for the Spec cPl in dogs with mild pancreatitis is likely a reflection of the stringent semiquantitative histopathologic scoring system that defined mild pancreatitis based on the presence of one or more lesion parameters affecting <10% of the surface area of the sections examined. It is plausible that the histopathologic definition of mild pancreatitis, as used in this study, is a broad one that includes many clinically normal animals. A second likely reason for the lower sensitivity of the Spec cPL observed in our study was the inclusion of pancreatic histopathologic features that are most consistent with chronic pancreatitis (lymphocytic inflammation, fibrosis, and atrophy) in our scoring system. A recent study of 14 cases of histopathologically confirmed chronic pancreatitis in dogs supports this hypothesis, as the sensitivity for the cPLI in that study ranged from 58% (95% CI 34–79%) to 26% (95% CI 10–51%) depending on the cutoff used. Our semiquantitative scoring system was selected to avoid biasing our results toward dogs with only acute pancreatitis (suppurative inflammation, pancreatic necrosis, peripancreatic fat necrosis, and edema). Our study did confirm, however, that the correlation coefficients for lymphocytic inflammation, fibrosis, atrophy, and hemorrhage were lower than that for suppurative inflammation and pancreatic/peripancreatic necrosis. Pancreatic leakage of lipase would be less likely to be associated with pancreatic fibrosis and atrophy; however, these correlations should be interpreted cautiously given that most of the variation in histopathologic scores cannot be explained by the serological tests. A correlation of 0.45 (suppurative inflammation) indicates that approximately 20% of the variation in histopathologic score can be explained by Spec cPL.
In a previous study of 22 dogs with macroscopic evidence of pancreatitis, a common macroscopic manifestation of pancreatitis was peripancreatic fat necrosis revealed by the presence of saponified fat. The most common histopathologic findings included neutrophilic pancreatic infiltration (19 of 22 dogs), pancreatic necrosis (16 of 22), and peripancreatic fat necrosis (18 of 22). That study also revealed the Spec cPL to have a higher correlation with the histopathologic pancreatitis activity index, a scoring system constructed for correlation with acute pancreatitis, compared to a histopathologic pancreatitis chronicity index. These combined findings suggest that the Spec cPL may be more useful for identifying dogs with acute pancreatitis compared to chronic pancreatitis.
All 12 dogs that had a Spec cPL > 400 μg/L and histopathologic evidence of mild pancreatitis had histopathologic evidence of a hepatopathy, and only one of these dogs had macroscopic evidence of pancreatitis. Likewise, 3 of 12 dogs that had a Spec cPL > 400 μg/L and histopathologic evidence of mild pancreatitis had histopathologic evidence of intestinal lesions (mild to moderate plasmacytic enteritis, B-cell lymphoma, or necrosuppurative serositis with crypt ectasia). Further studies are warranted to assess Spec cPL concentrations in dogs with extrapancreatic disease, in particular dogs with liver disease, given the findings of increased serum lipase activity in dogs with liver disease. Ultrasound evaluations were only performed in 8 of 70 dogs within 24 hours of euthanasia, and caution should be heeded in overinterpreting the findings given the limitations of a retrospective examination of still ultrasound images and the small number of dogs evaluated via ultrasound. Nevertheless, ultrasound evaluations did correctly diagnose pancreatitis that was ultimately confirmed via histopathology in all 8 dogs, and the Spec cPL was > 400 μg/L in 4 of these dogs and between 200 and 400 μg/L in 2 dogs. Two dogs with ultrasonographic evidence of pancreatitis and histopathologic evidence of mild and moderate pancreatitis had Spec cPLs of 77 and 76 μg/L, respectively.
The cTLI assay has been described as a highly sensitive and specific test for the assessment of exocrine pancreatic insufficiency, however, the performance characteristics of the assay are suboptimal for the diagnosis of pancreatitis, and the low sensitivity of this assay was confirmed in this study. Measurement of serum amylase activity was the least sensitive test for the diagnosis of mild or moderate to severe pancreatitis; however, the high specificity for this test and the cTLI can be explained by the relatively low numbers of dogs lacking histopathologic evidence of pancreatitis, and the fact that these tests were within the reference interval for all these dogs. Measurement of serum lipase activity was the second most sensitive serological marker for the diagnosis of pancreatitis, and the higher sensitivity ranging from 54 to 71% should be interpreted cautiously, given the nature of the study and the finding of a relatively large number of dogs with extrapancreatic disease.
A limitation of this study was that these dogs were necropsied for a variety of reasons, and that clinical inferences (weight loss, inappetance, vomiting, etc) that often are associated with pancreatitis could have been because of concurrent disease. The sensitivity data must be interpreted cautiously because of the relatively small number of dogs with histopathologic evidence of moderate to severe pancreatitis, and because of the broad inclusion of histopathologic features that are associated with both acute and chronic pancreatitis. Likewise, the relevance of mild pancreatic inflammation observed histopathologically may not reflect clinical disease, and the histopathologic findings of pancreatic suppurative inflammation, pancreatic necrosis, and peripancreatic fat necrosis may be more optimal histopathologic markers to assess pancreatitis.
In summary, the Spec cPL demonstrated the best overall performance characteristics (sensitivity and specificity) relative to the other serological tests for the diagnosis of pancreatitis in dogs, with pancreatitis being defined histopathologically as opposed to clinically. Caution should be heeded in interpreting the specificity results, because these were based on 7 normal pancreata only. The Spec cPL appears more suitable for the diagnosis of acute pancreatitis in dogs, although the test showed higher correlations with all of the pancreatic histopathologic features compared to the other serological tests. The sensitivity of the cTLI was inferior to that of serum lipase, underscoring the limitations of the cTLI for the diagnosis of canine pancreatitis. Further prospective studies are warranted to evaluate the diagnostic yield of combining abdominal ultrasound with determination of Spec cPL in dogs with suspected pancreatitis, in an effort to improve the combined sensitivities when both tests are performed in parallel.