Abstract
Feline pancreatitis is a commonly suspected illness and it has been proposed that some cases of feline pancreatitis may be caused by infection with Toxoplasma gondii or Bartonella species. Feline pancreatic lipase immunoreactivity (fPLI) is a test performed on serum that is commonly combined with other clinical findings as an indirect aid in the diagnosis of pancreatitis. The purpose of this study was to determine if there are associations between fPLI concentration and the presence of serum antibodies against T gondii or Bartonella species. Serum samples from 458 cats, for which serum fPLI concentrations had already been determined, were assayed by enzyme-linked immunosorbent assay for the presence of T gondii immunoglobulin (Ig) G (IgG) and IgM antibodies, and Bartonella species IgG antibodies. The association between fPLI concentration and T gondii or Bartonella species antibodies was determined. No statistically significant association was found between fPLI concentration and T gondii or Bartonella species antibodies, suggesting that serological tests for the organisms are not useful in cases with increased fPLI concentration.
Feline pancreatitis is a very commonly suspected illness in veterinary medicine. Affected cats most often present with vague, non-specific clinical signs including lethargy, anorexia, and dehydration. 1 Unlike dogs, cats with pancreatitis rarely vomit, and abdominal pain may go clinically undetected in ill cats. 2 The exact prevalence of feline pancreatitis is unknown, however one study reported histopathologic evidence of pancreatitis in greater than 60% of 115 cats presented for necropsy. 3 This suggests that many cases of pancreatitis may go undiagnosed, which is not surprising, given the difficulty in testing for the disease.
The gold standard for the diagnosis of feline pancreatitis is histopathology. However, obtaining pancreatic biopsy samples is impractical in many cases and is rarely done ante-mortem. 4 Abdominal radiographs are insensitive for the diagnosis of feline pancreatitis. 5 Abdominal ultrasound has variable sensitivity and specificity (sensitivity=62–80%, specificity=33–88%), which depend on the experience and skill level of the ultrasonographer, as well as the severity of the pancreatitis. 6 While the numbers of cases evaluated are small, non-invasive blood tests including serum trypsin-like immunoreactivity 5 and both serum lipase and amylase activities are believed to have poor diagnostic sensitivity and specificity for feline pancreatitis. 7 Serum pancreatic lipase immunoreactivity (PLI) is a relatively new test used to aid in the diagnosis of feline pancreatitis. One study showed the test to be both highly sensitive in cats with moderate to severe pancreatitis (100%) and highly specific for feline pancreatitis in healthy cats (100%). However, this study was limited by the small number of cats in which severe pancreatitis was diagnosed. The sensitivity of the feline PLI (fPLI) test is lower in cases of mild pancreatitis (54%) and the specificity is lower in sick cats with non-pancreatic illness (67%). Similar to other available non-invasive tests for feline pancreatitis, the fPLI test has limitations, and must be interpreted in conjunction with clinical signs, physical examination findings, and the results of additional testing for pancreatitis.
Trauma and feline infectious peritonitis are known causes of pancreatitis, and many clinicians suspect a relationship between inflammatory bowel disease, cholangiohepatitis, and pancreatitis. However, little is known about other potential causes of pancreatitis, and most cases of feline pancreatitis are considered to be idiopathic. It has been proposed that some cases of ‘idiopathic’ pancreatitis are actually the result of infection. 8 ‘Stealth’ organisms are organisms which are capable of escaping detection, either because they evade the immune system, cause subtle or non-specific clinical signs or laboratory abnormalities, or require special tests for identification. 9 Toxoplasma gondii and Bartonella species are common infections of cats, but most infected cats are normal and in those with clinical signs of disease, the diagnosis is not obvious based on clinical presentation, physical examination findings, or routine complete blood cell counts and serum biochemical analyses. 10 Thus, it is possible that organisms of this type may be responsible for some cases of feline pancreatitis, and may go undetected in some cases because specific diagnostic tests are not used in all cats.
Toxoplasma gondii is a tissue protozoan for which cats are the definitive host. Infected cats often present with vague, non-specific clinical signs including anorexia and depression. 11 T gondii can affect multiple organs and, therefore, can manifest as a variety of different syndromes, including fever, pneumonia, lymphadenopathy, uveitis, and hepatitis. 11 Organisms can be identified in tissue or feces. However, oocyst shedding in feces is transient, and tissue specimens are often unavailable. Thus, screening for toxoplasmosis is most often undertaken with serology; the T gondii seroprevalence in cats in the United States is greater than 30%.12,13 While most cats with T gondii antibodies are still infected, the presence of antibodies does not always correlate with active infection or the presence of disease. In contrast, most cats with T gondii in tissues have detectable serum antibodies and so seronegative cats are unlikely to have toxoplasmosis.
Bartonella species are Gram-negative, intracellular bacteria, which infect the red blood cells and tissues of cats. 14 The role of these organisms in feline illness remains controversial. However, there have been numerous speculations that infection with Bartonella species causes disease in cats. Some of the proposed illnesses include fever, uveitis, and stomatitis. Many asymptomatic cats have antibodies against Bartonella species and naturally infected cats commonly do not have any clinical signs. 15 However, experimentally infected cats have been reported to have numerous clinical signs, including fever, anemia, and lymphadenopathy, as well as histopathological changes. 15–17 Bartonella species are transmitted by the cat flea Ctenocephalides felis. Therefore, cats have a high risk of exposure. Seroprevalence rates range from 35 to 93%, depending on the characteristics and the geographical location of the population tested. 10,15,18–20 Results of serum antibody assays do not always correlate to presence of bacteremia; less than half of antibody result-positive cats may be actively infected and 3–15% of antibody result-negative cats may be bacteremic. 15 Thus, care must be taken when evaluating the results of Bartonella species antibody assays.
Both toxoplasmosis and bartonellosis have been proposed as causes of feline pancreatitis. In one study, 84% of 45 cats with confirmed toxoplasmosis, in which pancreatic biopsies were obtained, had histopathologic evidence of pancreatitis. 21 In studies in which small numbers of cats (<20) were experimentally infected with Bartonella species, the most common histopathologic changes noted were lymphoid hyperplasia, and lymphocytic or lymphoplasmacytic infiltration of numerous organs, including spleen, liver, lymph nodes, heart, and kidney.16,17 The pancreas was not specifically evaluated in those studies. The most common histopathologic change reported with chronic pancreatitis in cats is lymphocytic inflammation. 3 These findings support the theory that some cases of pancreatitis in cats may be caused by Bartonella species or T gondii infections. Veterinarians are using the fPLI assay as an aid in the diagnosis of pancreatitis and there is anecdotal information that veterinarians are being advised to test cats with suspected pancreatitis for Bartonella species or T gondii antibodies. Thus, the purpose of this study was to determine if there is an association between fPLI concentrations and the presence of serum antibodies against T gondii or Bartonella species.
Methods and materials
Samples
A total of 464 serum samples were obtained from the Gastroenterology Laboratory at Texas A&M University (http://www.cvm.tamu.edu/gilab). The sera came from cats for which the referring veterinarian had submitted the samples to be assayed for fPLI. The fPLI concentration was determined at that time and the samples stored at −80°C until used for this study. An fPLI concentration of 2.0–6.8 μg/l is considered normal, while a serum fPLI concentration >12 μg/l is elevated and is consistent with pancreatitis. Concentrations of fPLI between 6.9 and 12 μg/l are considered to be in the questionable range and may suggest pancreatitis. Samples were selected sequentially, and based solely on availability, and selected samples were shipped by overnight express on cold packs to Colorado State University where they were stored at −80°C. In addition to the fPLI concentration, information available for each sample included the age, breed, and state of residence of the cat. The state of origin was classified as low flea risk (Alaska, Arizona, Colorado, Idaho, Montana, Nevada, New Mexico, Utah, Wyoming) or high flea risk (all other states) based on a previous study. 19 Samples with fPLI concentrations <2 μg/l were excluded because they were below the lower limit of the reference range, leaving 458 samples for analysis.
Bartonella species and T gondii serology
The sera were thawed at room temperature and assayed the same day for T gondii IgM, T gondii IgG, and Bartonella species IgG by enzyme-linked immunosorbent assays (ELISAs) using the standard operating procedures of the Endocrinology/Specialized Infectious Diseases Laboratory at Colorado State University (http://www.dlab.colostate.edu/webdocs/general/off1.html). ELISA results of ≥1:64 were considered positive for each of the respective antibodies.
Statistical analysis
To determine the association between fPLI concentration and T gondii or Bartonella species antibodies, samples were first classified, according to fPLI concentration, as normal (fPLI=2.0–6.8 μg/l), questionable (fPLI=6.9–12 μg/l), or positive for pancreatitis (fPLI>12 μg/l). Statistical analysis was performed using a Fisher's exact test, to determine the relationship between serum fPLI status (normal, questionable, or positive) and the presence of T gondii or Bartonella species antibodies. In addition, the relationship between the presence of T gondii or Bartonella species antibodies and fPLI status was evaluated by logistic regression, in a model which included fPLI status, age, and the interaction between these factors in the statistical model. The fPLI was also evaluated in this model as a continuous variable. Due to the small number of cats from non-flea states, this variable was not included in the final model. Odds ratios and 95% confidence intervals (CIs) were calculated for each relationship. Significance was defined as P<0.05.
Results
A total of 458 samples were included in the study. Samples came from cats from 43 states and Washington, DC. Cats ranged from 1 month to 22 years in age, with a median age of 11 years. The sample distribution included 376 domestic shorthair cats and 82 purebred cats. Breeds represented included Siamese (21 cats), Persian (14 cats), Maine Coon (13 cats), Himalayan (nine cats), Abyssinian (four cats), Birman (three cats), Burmese (three cats), Balinese (two cats), Bengal (two cats), Manx (two cats), Ocicat (two cats), Tonkinese (two cats), Oriental Shorthair (one cat), Rag Doll (one cat), Rex (one cat), and American Shorthair (one cat). One sample was from a cat of unknown breed. Samples came from the following states; Arizona (seven cats), California (93 cats), Colorado (11 cats), Connecticut (10 cats), Florida (37 cats), Georgia (six cats), Hawaii (one cat), Idaho (one cat), Illinois (33 cats), Indiana (two cats), Iowa (five cats), Kansas (four cats), Kentucky (four cats), Louisiana (five cats), Maine (four cats), Maryland (13 cats), Massachusetts (23 cats), Michigan (19 cats), Minnesota (10 cats), Mississippi (two cats), Missouri (two cats), Montana (three cats), Nebraska (two cats), Nevada (four cats), New Hampshire (seven cats), New Jersey (15 cats), New Mexico (two cats), New York (21 cats), North Carolina (three cats), North Dakota (one cat), Ohio (nine cats), Oklahoma (one cat), Oregon (six cats), Pennsylvania (18 cats), Rhode Island (one cat), South Carolina (three cats), Tennessee (one cat), Texas (26 cats), Vermont (one cat), Virginia (20 cats), Washington (eight cats), Wisconsin (11 cats), Wyoming (one cat), and Washington DC (two cats). Of the samples, 432/458 (94.3%) came from cats living in high flea-risk states.
A total of 179 (39.1%) cats had an fPLI concentration greater than 12 μg/l, which is considered by the laboratory as consistent with pancreatitis. Of these cats, 151 (84.3%) were domestic shorthair cats, and 28 (15.6%) were purebred; 167 (93.3%) were from high flea-risk states, and 12 (6.7%) were from low flea-risk states. The fPLI positive cats ranged in age from 2 to 20 years, with a median age of 13 years. A total of 104 (22.7%) cats had fPLI concentrations which were elevated, but in the questionable range (6.9–12 μg/l). Of these, 82 (78.8%) were domestic shorthair cats, and 22 (21.2%) were purebred; 101 (97.1%) came from high flea-risk states, and three (2.9%) came from low flea-risk states. Cats with fPLI concentrations in the questionable range ranged in age from 1 to 20 years, with a median age of 11 years. Of the 175 (38.2%) cats which had fPLI concentrations in the normal range (2.0–6.8 μg/l), 143 (81.7%) were domestic shorthair cats, and 32 (18.3%) were purebred; 164 (93.7%) came from high flea-risk states, and 11 (6.3%) came from low flea-risk states. These cats ranged in age from 1 month to 22 years, with a median age of 8 years.
The T gondii immunoglobulin (Ig) M IgM, T gondii IgG, and Bartonella species IgG seroprevalence rates, stratified by fPLI concentration ranges, are listed in Table 1. Fisher's exact test indicated no statistically significant relationship between fPLI category and the positive rate of T gondii IgM, T gondii IgG, or Bartonella species IgG antibodies. Logistic regression analysis including fPLI as a categorical outcome indicated that cats with fPLI concentrations categorized as either elevated or in the questionable range were no more likely to be seropositive for T gondii IgM (odds ratio=0.62, 95%CI=0.23, 1.64, P=0.33), T gondii IgG (odds ratio=0.77, 95%CI=0.25, 2.35, P=0.65), or Bartonella species IgG antibodies (odds ratio=0.96, 95%CI=0.46, 2.02, P=0.91), than cats with fPLI concentrations in the normal range. These relationships were not influenced by age. No statistically significant association was found between the actual fPLI concentration and T gondii IgM antibodies (odds ratio=1.00, 95% CI=0.97, 1.03, P=0.94), T gondii IgG antibodies (odds ratio=1.00, 95%CI=0.99, 1.01, P=0.96), or Bartonella species IgG antibodies (odds ratio=1.00, 95%CI=0.99, 1.01, P=0.59). There was no influence of age on these relationships.
Seroprevalence rates of T gondii IgG, T gondii IgM, and Bartonella species IgG in cats with normal, questionable, and elevated serum fPLI concentrations
Discussion
The results of this study indicate that cats in the study population with elevated fPLI concentrations are no more likely to have serum antibodies against T gondii or Bartonella species than cats with fPLI concentrations within the reference range. These results suggest that the performance of T gondii or Bartonella species antibody tests on all cats with suspected or confirmed pancreatitis is likely to have little clinical utility. Our study is limited by our sampling method. Serum samples had been submitted for the fPLI assay prior to the initiation of our study. Information regarding clinical signs and physical examination findings was not available, and the reasons for sample submission were unknown. Samples may have been submitted for reasons other than suspicion of pancreatitis, thereby biasing our results. The results of this study do not exclude the possibility that pancreatitis is caused by toxoplasmosis or bartonellosis. This study assessed the relationship between fPLI concentration and the presence of antibodies against these infectious agents. However, the gold standard for diagnosis of pancreatitis is histopathology rather than fPLI concentration which may have influenced the results of the study. 22 Additionally, the presence of T gondii or Bartonella species antibodies does not prove active infection and the absence of antibodies does not always exclude the possibility of disease. 23
It has been found that cats with experimental Bartonella species infections commonly have lymphocytic inflammation in numerous organ systems, and it is known that lymphocytic inflammation is the most common histopathologic finding in cats with chronic pancreatitis.3,16,17 It is tempting to speculate that some cases of feline chronic pancreatitis might therefore be caused by Bartonella species infections. However, mild lymphocytic inflammation in the pancreas may not cause clinical signs and could be present in cats without causing an elevation in fPLI concentration. If so, a relationship between pancreatitis and Bartonella species infection might go unidentified in a study which looks for an association between fPLI concentration and Bartonella species antibodies. Therefore, while this study provides new information about the relationships between serum fPLI and serologic test results for these agents, further studies are needed to truly define the roles of T gondii and Bartonella species in feline pancreatitis. The relationships between pancreatitis and T gondii or Bartonella species infections might be better defined by looking for an association between histopathologic evidence of pancreatitis and the presence of organisms in tissue or polymerase chain reaction (PCR) evidence of T gondii or Bartonella species DNA.
The overall seroprevalences of T gondii and Bartonella species antibodies in the population included in our study were 16.2% (74/458) and 19.0% (87/458), respectively. These seroprevalence rates are lower than the previously reported seroprevalence rates of 31.6% and 27–93% for T gondii and Bartonella species, respectively. 10,13,18–20 The discrepancy between previous reports and the results of our study might be explained by sample bias, given that the samples were submitted to the diagnostic laboratory for a purpose other than this study. All of the samples included in this study came from client-owned cats, whereas previous studies reporting seroprevalence rates have included feral populations. Client-owned cats are likely to have a decreased risk of exposure to T gondii or Bartonella species, and possibly even to pancreatitis, when compared to a general population of cats which may include feral cats that may not be receiving optimal diets, are hunting, and are not being administered flea control products.
Previous studies have suggested that pancreatitis may be a common syndrome in cats with toxoplasmosis. One study reported that 84% of 45 cats with confirmed toxoplasmosis, which had pancreatic biopsies performed, had histopathologic evidence of pancreatitis. 21 The results of our study did not provide support for an association between elevated PLI concentrations and T gondii serum antibodies. However, our study examined a population of cats suspected of having pancreatitis, and determined if there was an association with T gondii antibodies rather than examining a population of cats with suspected toxoplasmosis and determining if there was an association with elevations in fPLI. Pancreatitis is multi-modal and likely has many causes. In selecting for a population of cats suspected of having a disease with multiple causes, we may have diluted out the population of cats infected with toxoplasmosis. It is possible that cats with toxoplasmosis commonly have pancreatitis, but that, due to the many other causes of pancreatitis, toxoplasmosis is not a common cause of pancreatitis overall.
In conclusion, we failed to find an association between elevated fPLI concentrations and antibodies against T gondii or Bartonella species. This information is clinically useful because it suggests that serology for T gondii and Bartonella species are not essential components of the diagnostic evaluation of all cats suspected of having pancreatitis based on an elevated fPLI. However, additional studies are needed to more accurately define the true relationships between these diseases.
Footnotes
Acknowledgments
Funding for the T gondii and Bartonella species serological tests was provided by the Center for Companion Animal Studies at Colorado State University.