Pancreatitis is a common disease in cats that is difficult to diagnose.
Pancreatitis is a common disease in cats that is difficult to diagnose.
To determine the sensitivity and specificity of ultrasonographic changes of the pancreas with serum feline pancreatic lipase immunoreactivity (fPLI) as the standard for diagnosis of pancreatitis.
35 cats with clinical signs consistent with pancreatitis with an abdominal ultrasound examination and serum fPLI concentration measured within 3 days of the ultrasound.
Retrospective study: Pancreatic thickness, pancreatic margination, pancreatic echogenicity, and peripancreatic fat echogenicity were evaluated. Sensitivity and specificity were calculated with an elevated serum fPLI concentration indicative of pancreatitis as the standard for diagnosis.
Serum fPLI was elevated and diagnostic for pancreatitis in 19 of 35 cats. The single ultrasound characteristic with the highest sensitivity was hyperechoic peripancreatic fat at 68% (95% confidence interval = 44–87%), indicating a moderate probability that cats with pancreatitis will have this abnormality on ultrasonographic examination. Specificity was >90% for each of increased pancreatic thickness, abnormal pancreatic margin, and hyperechoic peripancreatic fat. The sensitivity and specificity of ultrasound were 84% (95% confidence interval = 60–97%) and 75% (95% confidence interval = 48–93%), respectively, in cats with elevated serum fPLI indicative of pancreatitis.
The presence of a thick left limb of the pancreas, severely irregular pancreatic margins, and hyperechoic peripancreatic fat in cats with appropriate clinical signs and elevated serum fPLI are highly supportive of pancreatitis.
feline pancreatic lipase immunoreactivity
inflammatory bowel disease
Pancreatitis is a common disease in cats, yet antemortem diagnosis remains challenging. Pancreatic disease often produces nonspecific clinical signs such as lethargy, vomiting, and anorexia or remains subclinical.[1-10] Definitive diagnosis requires biopsy, which is not feasible in all clinical cases. In addition, the presence of inflammation on histopathology does not necessarily correspond with clinical signs of illness. Pancreatitis in cats, whether acute or chronic, has been shown to be a multifocal disease histologically, which further confounds diagnosis. Healthy cats often have histologic lesions consistent with chronic pancreatitis (45%; 18 of 41).
In part, the challenge of antemortem diagnosis of pancreatitis in cats is due to the lack of a reliable noninvasive diagnostic test for the disease. Histopathology is currently considered the gold standard for diagnosis, but noninvasive methods of diagnosis have recently been evaluated. Abdominal ultrasound has a low sensitivity, ranging from 11–67%, with a specificity that is consistently higher.[5-7, 12] Ultrasound has a sensitivity of 80% in cases of moderate to severe pancreatitis and 62% in cases of mild pancreatitis with a specificity of 88%. The variation in sensitivity is likely related to the experience level of the ultrasonographer, type of equipment, severity of lesions, and lack of standardized diagnostic criteria.
Ultrasonographic changes associated with pancreatitis include a hypoechoic pancreas, hyperechoic surrounding mesentery and fat, abdominal effusion, pancreatic enlargement, pancreatic pseudocysts, pancreatic calcification, and a corrugated duodenum.[5-7, 13] A hyperechoic pancreas, indicative of fibrosis, has also been reported. It is important to note that ultrasonographic changes of the pancreas because of various disorders such as pancreatitis, pancreatic neoplasia, and nodular hyperplasia overlap, and the presence of an ultrasonographically normal pancreas do not rule out pancreatic disease.
Development of the species-specific immunoassay that measures serum pancreatic lipase (feline pancreatic lipase immunoreactivity or fPLI) has the potential to improve the clinical approach to pancreatitis in cats. Serum fPLI concentration has been shown to have a sensitivity of 100% in 5 cats with moderate to severe pancreatitis, whereas that of ultrasound was 80%. The sensitivity for mild pancreatitis was lower at 54% with a resultant overall sensitivity of 67%. There is no universally accepted single noninvasive diagnostic test for pancreatitis in cats, but serum fPLI appears to be the most sensitive test currently available.
The aim of this retrospective study was to determine the sensitivity and specificity of abdominal ultrasound with serum fPLI as a standard for diagnosis of pancreatitis in cats. Ultrasonographic changes of the pancreas most often associated with pancreatitis were selected for comparison with serum fPLI concentration.
Medical records from the Virginia-Maryland Regional College of Veterinary Medicine Veterinary Teaching Hospital of cats in which a serum fPLI concentration and abdominal ultrasound examination had been performed were reviewed. Cases were included if at least 2 clinical signs consistent with pancreatitis, including anorexia, lethargy, weight loss, vomiting, diarrhea, icterus, and abdominal pain were recorded, ultrasound images were available for review retrospectively, and the fPLI was measured in serum within 3 days of the ultrasound examination.
Feline pancreatic lipase immunoreactivity was analyzed via radioimmunoassay (RIA) in the majority of cases.1, Over the course of the study (August 2003–April 2009), the reference interval for serum fPLI concentration analyzed via RIA was reestablished twice because of the use of new batches of antibodies employed in the RIA. Four cases after October 2008 had serum fPLI assayed with an ELISA test2 that has been stated to have a 98% diagnostic agreement with the fPLI-RIA and a different range than the RIA. Serum fPLI concentrations were assessed as either within the respective reference intervals or elevated in the range established by the laboratory to be “consistent with pancreatitis.” The reference intervals and concentrations considered indicative of pancreatitis were 1.2–3.8 and >8 μg/L, 2.6–6.8 and >12 μg/L, 4.1–12.9 and >12.9 μg/L for the RIA and <3.5 and >5.3 μg/L for the ELISA, respectively.[9, 12, 15],3 Cats with a serum fPLI concentration that fell between the upper limit of the reference interval and that consistent with pancreatitis were excluded.
All ultrasound examinations were performed by a board-certified radiologist (MML) with cats in dorsal recumbency using 8.5 MHz sector and 14 MHz linear array transducers.4 The body of the pancreas was identified directly caudal to the pyloroduodenal angle, ventral to the portal vein, and medial to the proximal duodenum. The left limb of the pancreas was identified caudal to the stomach, medial to the spleen, and cranial to the left kidney. The right limb was identified dorsomedial to the descending duodenum, but it was not consistently seen in every cat. Therefore, measurements were confined to the left limb of the pancreas.
Still ultrasound images were evaluated retrospectively by a single author (MML) for the following characteristics: (1) pancreatic thickness, (2) pancreatic margination, (3) pancreatic echogenicity, and (4) echogenicity of the peripancreatic fat (Fig 1).[5, 13, 18, 19] The reviewer was unaware of the serum fPLI concentration at the time of retrospective analysis, and no reference was made to the original imaging reports for each individual ultrasound examination.
The thickness of the pancreas, or maximum width in the ventrodorsal dimension, was measured only on the left limb in centimeters with electronic calipers. All measurements were made in a longitudinal image plane (short axis to the body), as transverse images were not available for all examinations. Measurements were taken from the best-defined image for each pancreas with measurements >1 cm considered abnormal.[13, 20]
Pancreatic margination was subjectively divided into three grades where Grade 1 corresponded to a smooth and linear pancreatic margin. Grade 2 corresponded to a pancreas with mild undulations in its margin, and severe irregularity to the pancreatic margin was classified as Grade 3. Normal pancreatic echogenicity was considered to be isoechoic to the adjacent liver. A pancreas hypoechoic to the liver was classified as abnormal.
The peripancreatic fat was considered hyperechoic when there was marked contrast between the pancreas and adjacent peripancreatic fat. If hyperechoic fat was identified throughout the abdomen, consistent with peritonitis, the cat was excluded from the study.
Sensitivity and specificity were determined for each ultrasound finding in all cats with an elevated serum fPLI indicative of pancreatitis as the standard for diagnosis. 95% confidence intervals for sensitivity and specificity were calculated on the basis of a binomial distribution. Statistical analyses were performed by commercially available software.5
Medical records of 40 cats were identified where a serum fPLI concentration, full abdominal ultrasound examination, and clinical signs consistent with a diagnosis of pancreatitis were recorded. Five cats were excluded from analysis because the serum fPLI concentration was between the upper limit of the reference interval and that consistent with pancreatitis, leaving 35 cats that were included in the study. Ages ranged between 5 months and 18 years with a median age of 11 years. There were 22 neutered males, 1 intact male, and 12 spayed females. The breed distribution was 23 domestic short hair, 5 domestic long hair, 5 Siamese, and 1 each Burmese and Maine Coon.
The most common clinical signs recorded were weight loss (14 of 35; 40%), anorexia (13 of 35; 37%), chronic vomiting (9 of 35; 26%), lethargy (6 of 35; 17%), and icterus (4 of 35; 11%). The clinical diagnosis listed in the medical record for cats with a normal serum fPLI included chronic renal failure (1), pancreatitis (1), pancreatic carcinoma (1), idiopathic megaesophagus (1), cholelithiasis (1), inflammatory bowel disease (IBD) (1), gastrointestinal lymphoma (1), hepatic lipidosis (2), cholangiohepatitis (3), and open (4). Diagnoses listed in cats with elevated serum fPLI, excluding pancreatitis, were diabetes mellitus (1), diabetic ketoacidosis (1), gastritis (1), gastrointestinal carcinoma (1), hepatocellular carcinoma and cholangiocarcinoma (1), extrahepatic biliary obstruction and cholangiohepatitis (2), chronic renal failure (2), cholecystitis (3), and IBD (3). Six cats with an elevated serum fPLI concentration had concurrent diseases. These included combinations of pancreatitis, IBD, chronic renal failure, cholecystitis, or cholangiohepatitis. One cat had IBD and cholecystitis without pancreatitis.
Sixteen (46%) cats had serum fPLI concentrations within the reference interval (Table 1). The 5 cats ultimately diagnosed with pancreatitis on the basis of histopathology had elevated serum fPLI concentrations [14.5 μg/L (>5.3 μg/L); 30.6–79.4 μg/L (>12 μg/L); 90.1–129 μg/L (>12.9 μg/L].
|Any US Abnormality||Pancreatic Thickness||Pancreatic Margin||Pancreatic Echogenicity||Peripancreatic Fat Echogenicity|
|Absent||Present||Normal||Inc||Normal||Grade 2||Grade 3||Normal||Hypo||Normal||Hyper|
|Normal serum fPLI||12||4||15||1||13||2||1||14||2||16||0|
|Abnormal serum fPLI||3||16||12||7||7||4||8||10||9||6||13|
Increased pancreatic thickness was present in 7 of the 19 cats (37%) with an elevated serum fPLI concentration (Table 1). One of the cats with abnormal pancreatic thickness had chronic pancreatitis confirmed via histopathology. Abnormal pancreatic margin was present in 12 (63%) cats with an elevated serum fPLI concentration and in 3 cats with a normal serum fPLI concentration. A Grade 2 pancreatic margin was identified in 4 (21%) cats with an elevated serum fPLI concentration and in 2 cats with a normal value. A Grade 3 pancreatic margin was identified in 8 cats (42%) with an elevated serum fPLI concentration, two of which had acute and chronic pancreatitis, respectively, noted on histopathology. A Grade 3 pancreatic margin and normal serum fPLI concentration were identified in 1 cat. A hypoechoic pancreas was found in 9 (47%) cats with a serum fPLI concentration above the reference interval and in 2 cats with a normal serum fPLI concentration. Two cats with confirmed chronic pancreatitis had hypoechoic pancreata. Thirteen (68%) cats had hyperechoic peripancreatic fat and an elevated serum fPLI concentration. No cats with a normal serum fPLI concentration had this change. All 5 cats with histopathologically confirmed pancreatitis exhibited hyperechoic peripancreatic fat.
Ultrasonographic findings with the highest sensitivity for an elevated serum fPLI concentration were Grade 3 pancreatic margination and hyperechoic peripancreatic fat, indicating that these characteristics were more likely to be found in cats with pancreatitis on the basis of an elevated serum fPLI concentration. Increased pancreatic thickness and a Grade 2 pancreatic margination were less sensitive (Table 2). Specificity of ultrasonographic abnormalities was very high with a thickened pancreas, Grade 3 pancreatic margination, and hyperechoic peripancreatic fat having greater than 90% specificity for elevated serum fPLI concentration. These ultrasound characteristics were observed less frequently in cats with a normal serum fPLI concentration. The sensitivity and specificity for any of the 4 ultrasonographic abnormalities were 84% (95% confidence interval = 60.4–96.6%) and 75% (95% confidence interval = 47.6–92.7%), respectively, in all cats with elevated serum fPLI concentration.
|Any US Abnormality||Pancreatic Thickness||Pancreatic Margin||Pancreatic Echogenicity||Peripancreatic Fat Echogenicity|
|Grade 2||Grade 3|
|Sensitivity (%) (95% CI)||84.2 (60.4, 96.6)||36.8 (16.3, 61.6)||36.4 (10.9, 69.2)||53.3 (26.6, 78.7)||47.4 (24.5, 71.1)||68.4 (43.5, 87.4)|
|Specificity (%) (95% CI)||75 (47.6, 92.7)||93.8 (69.8, 99.8)||86.7 (59.5, 98.3)||92.9 (66.1, 99.8)||87.5 (61.7, 98.5)||100 (79.4, 100)|
Other significant ultrasonographic findings in the entire population of cats included an enlarged, hyperechoic liver in 6 (17%), changes in hepatic echotexture in 3 (8%), gall bladder distension in 7 (20%), gall bladder wall thickening in 11 (31%), choleliths in 2 (5%), and a thickened, tortuous common bile duct in 12 (34%). A thickened muscularis layer of the small intestine was present in 10 of the cats (29%). Mild to moderate mesenteric lymph node enlargement and the presence of free peritoneal fluid were present each in 6 (17%) cats. Four of the 5 cats with histopathologically confirmed pancreatitis had a thickened gall bladder wall and tortuous common bile duct. Concurrent cholangiohepatitis, cholangitis, or cholecystitis were confirmed in these cats on necropsy.
Fourteen cats had histopathology of various tissues available either via postmortem examination (9) or endoscopic biopsies (5). Seven of these cats had a serum fPLI concentration within the reference interval. The only cat in this group in which the pancreas was evaluated histologically had a normal result. The final histopathologic diagnosis for cats with serum fPLI concentrations within the reference interval was steroid hepatopathy (1), IBD (2), hepatic lipidosis (1), cholangiohepatitis (1), and normal duodenum (2).
Seven cats with a serum fPLI concentration above the reference interval had histopathology of 1 or more tissues performed. Five cats were definitively diagnosed with acute or chronic pancreatitis on the basis of histopathology as well as with concurrent diseases. One cat had acute pancreatitis with multifocal suppurative inflammation of the pancreas and cholangiohepatitis. Another cat had mild chronic multifocal lymphoplasmacytic inflammation and cholecystitis. A 3rd cat had evidence of acute and chronic pancreatitis with mild chronic changes, fibrosis, and areas of lymphocytic and neutrophilic inflammation as well as cholangiohepatitis. Mild and severe fibrosis consistent with chronic pancreatitis was noted in 2 cats, respectively. The cat with mild fibrosis also had hepatocellular carcinoma and cholangiocarcinoma. The cat with severe fibrosis had both cholangiohepatitis and cholecystitis. Only one of the remaining 2 cats had histopathology of the pancreas available, and the results were normal. The final histopathologic diagnosis for the remaining 2 cats with an elevated serum fPLI concentration was gastrointestinal carcinoma (1) and normal duodenum (1).
The results of this study, using an elevated serum fPLI concentration as a marker of pancreatitis in cats with compatible clinical findings, determined that ultrasound has moderate sensitivity for detecting pancreatitis as indicated by elevated serum fPLI concentration. Increased echogenicity of the mesenteric fat immediately surrounding the pancreas had the highest sensitivity at 68%. Sensitivity of ultrasound for pancreatitis was higher in this study (84% for any ultrasonographic abnormality) than others, where the sensitivity of pancreatic ultrasound for cats with histopathologically confirmed pancreatitis has ranged from 11–67%. In another study of cats with elevated serum fPLI but without pancreatic histopathology, the sensitivity was 25%. The high sensitivity reported in this study is most likely because of the fact that this population consisted only of cats with a clinical suspicion of pancreatitis. It is possible that cats in this study had moderate to severe pancreatitis, prompting measurement of serum fPLI which was a criterion for inclusion in the study. The sensitivity of ultrasonography has been shown to increase with increasing severity of pancreatitis as determined histopathologically. Similar to other studies, the specificity of certain ultrasonographic abnormalities of the pancreas was very high.[5-7, 12] Therefore, finding any of the abnormalities described in this study, particularly increased thickness of the pancreas or hyperechoic peripancreatic fat, proves to be an excellent marker for pancreatitis in cats with appropriate clinical signs.
Definitive diagnosis of pancreatitis is dependent on demonstrating inflammation or fibrosis on histopathology. However, a histopathologic diagnosis of pancreatitis does not necessarily correlate with clinical signs of the disease. Pancreatitis was identified on necropsy of 67% of 115 cats, including 45% of 41 clinically normal cats. In addition, inflammation is most frequently focal or multifocal rather than diffuse, and if a pancreatic biopsy specimen is obtained from an unaffected area, a lesion could be overlooked. Because of these factors and the invasiveness of pancreatic biopsy, an elevated serum fPLI concentration in cats with other clinical findings consistent with pancreatitis was chosen as a proxy for histopathologic confirmation. It is difficult to directly compare the sensitivities between studies that use 2 different gold standards for the diagnosis of pancreatitis. Serum fPLI has a sensitivity of 100% for moderate to severe pancreatitis on the basis of a semiquantitative histopathology grading scale established to reflect the severity of lesions. Pancreatic tissue was acquired in only 6 of the 35 cats reported here, and a standard protocol for biopsy and histologic assessment was not in place to ensure that multiple sections were evaluated. A single pancreatic biopsy specimen is generally not sufficient to exclude a diagnosis of pancreatitis. In light of its multifocal nature, a diagnosis of pancreatitis may have been missed, particularly in the 1 cat whose pancreas was normal on histopathologic evaluation. Thus, the 84% sensitivity of ultrasound reported in this study should be interpreted with caution.
Specificity was relatively high for all ultrasonographic findings. Hyperechoic peripancreatic fat produced a specificity of 100% in this study, in comparison with 73% based on the gold standard of histopathology. Mild irregularity of the pancreatic margin (Grade 2) proved to be the least-specific ultrasound abnormality. Evaluation of the specificity of serum fPLI concentration for diagnosis of pancreatitis is limited but reported to be 91% in a population of normal cats and cats with clinical signs of pancreatitis without histologic evidence of the disease. Similar to sensitivity in this study, the higher specificity may be due, in part, to inclusion of cats with a clinical suspicion of pancreatitis compared with a population that included apparently healthy cats.[12, 18]
Serum fPLI concentration is considered an imperfect gold standard although it has been identified as a marker with the highest sensitivity and specificity available. The RIA has been reported to be sensitive, accurate, and precise. Peer-reviewed information on the development and validation of the ELISA test,2 which has replaced the RIA, remains unpublished. However, the manufacturer reports a 92% agreement between the RIA and ELISA with values within the reference interval. Agreement decreases to 82% at values greater than the reference interval. A recent study reported a 79% sensitivity and 82% specificity of the ELISA test in both healthy cats and cats with clinical signs consistent with pancreatitis.3 On the basis of a test population of 182 cats, these authors concluded that the ELISA test was an appropriate test to screen for pancreatitis.
Other limitations of serum fPLI include a lack of evaluation of its stability, half-life, and biologic variability as well as a lack of evaluation of diagnostic performance in cats in the face of concurrent diseases. In this study, selection of a maximum of 3 days between measurement of serum fPLI and abdominal ultrasound examination was arbitrary. Without knowledge of the biologic half-life of circulating fPLI, the dynamics of its release from the pancreas during an episode of pancreatitis, or the duration and severity of pancreatitis in cats of this study, it is possible that ultrasound abnormalities may have resolved in the period between sampling for serum fPLI and ultrasound. If this were the case, ultrasound of the pancreas might be more sensitive than indicated in this study.
Clinical signs and physical examination findings for the cats included in this study were similar to those previously reported in cases where pancreatitis was included on the list of differential diagnoses.[3, 6, 8, 12, 22] Concurrent diseases were diagnosed in cats with an elevated serum fPLI concentrations in this study. These included gastrointestinal lymphoma, IBD, hepatic lipidosis, and cholangiohepatitis. Concurrent IBD, hepatobiliary disease, and pancreatitis have been previously noted, although their relationship is unclear.[4, 6, 12, 18, 23-25] Although serum fPLI is reported to be specific for pancreatitis, in a study of 23 cats with histopathologically confirmed IBD, 69% had an increased serum fPLI concentration. Pancreatic biopsies were not available in that study, and only 4 of 22 cats with an abdominal ultrasonographic evaluation had pancreatic changes consistent with possible pancreatitis. In this study, 1 cat with a final histopathologic diagnosis (gastrointestinal lymphoma and cholangiohepatitis), elevated serum fPLI concentration, and histopathology of the pancreas did not have pancreatitis. Although this patient was necropsied, the number of samples of the pancreas that were evaluated is unknown. It is possible that focal pancreatitis may have been missed on histopathologic evaluation. Thus, the number of clinically ill cats without pancreatitis that have been evaluated is small, and the influence of concurrent disease is unknown. With the exception of the 5 confirmed cases of pancreatitis, we are unable to determine the contribution of pancreatitis to the clinical signs of the cats in this study.
Limitations of this study include its retrospective nature, particularly with respect to the analysis of still ultrasound images for subjective measurements. Although the same radiologist performed the real-time evaluations as well as the retrospective analysis, subtle changes in tissue echogenicity are often best evaluated in real time. The retrospective analysis occurred up to 6 years after the original ultrasound examination, and no reference to previous ultrasound reports was conducted. Another study limitation includes the lack of ultrasonographic evaluation of the right limb of the pancreas. The right limb was inconsistently visualized in this study with ultrasound; thus, analysis was limited to the left limb. The right limb of the pancreas is reported to be thinner than the left limb, and duodenal and colonic gas are more likely to prevent complete visualization of this portion of the pancreas. Based on the multifocal nature of pancreatitis in cats, both ultrasonographic characteristics and histopathologic findings associated with pancreatitis of the right limb may have been missed.
The high specificity of hyperechoic peripancreatic fat on abdominal ultrasound examination in cats with appropriate clinical signs indicates that this finding will be of considerable utility in the diagnosis of pancreatitis. However, use of elevated serum fPLI concentration as the benchmark for diagnosis of pancreatitis may have reduced the validity of these findings. Further study will be necessary to determine the place of abdominal ultrasonography in the diagnosis of pancreatitis in cats.
Conflict of Interest: Authors disclose no conflict of interest.
Spec fPLI™, Gastrointestinal Laboratory, Texas A&M University, College Station, TX
Spec fPLI® Test, IDEXX Laboratories, Westbrook, ME
Forman, MA, Shiroma, J, Armstrong, PJ, et al. Evaluation of feline pancreas-specific lipase (Spec fPL®) for the diagnosis of feline pancreatitis. 2009 ACVIM Forum. J Vet Intern Med 2009;23:733–734 (abstract)
ACUSON Sequoia™ 512, Siemens Medical Solutions USA, Inc, Mountain View, CA
SAS® 9.3, SAS Institute, Inc, Cary, NC