Comparative analysis of expert and machine-learning methods for classification of body cavity effusions in companion animals.

A rule-based expert system using CLIPS programming language was created to classify body cavity effusions as transudates, modified transudates, exudates, chylous, and hemorrhagic effusions. The diagnostic accuracy of the rule-based system was compared with that produced by 2 machine-learning methods: Rosetta, a rough sets algorithm and RIPPER, a rule-induction method. Results of 508 body cavity fluid analyses (canine, feline, equine) obtained from the University of California-Davis Veterinary Medical Teaching Hospital computerized patient database were used to test CLIPS and to test and train RIPPER and Rosetta. The CLIPS system, using 17 rules, achieved an accuracy of 93.5% compared with pathologist consensus diagnoses. Rosetta accurately classified 91% of effusions by using 5,479 rules. RIPPER achieved the greatest accuracy (95.5%) using only 10 rules. When the original rules of the CLIPS application were replaced with those of RIPPER, the accuracy rates were identical. These results suggest that both rule-based expert systems and machine-learning methods hold promise for the preliminary classification of body fluids in the clinical laboratory.     

EVALUATION OF THE ABDOMEN USING ULTRASOUND FOLLOWING A VENTRAL MIDLINE CELIOTOMY

Over the last 20 years, ultrasound has become a useful diagnostic tool in veterinary medicine. The purpose of this study was to use ultrasonographic examinations to determine if intra-abdominal contents could be visualized. Nineteen dogs, ranging from 8 months to 18 years of age were included. Each animal was evaluated sonographically prior to surgery, 24 h after surgery, and then 2-6 weeks following the surgical procedure. The results of our study suggest that ultrasound is an effective means for evaluation of the abdomen and surgery site following surgery. Both the abdomen and surgery site were clearly visualized in all dogs. In addition, subjective evaluation of the intra-abdominal contents was performed. The limiting factor in this study was not the presence of air, as was expected. Rather the limiting factor for appropriate visualization was pain. The information obtained could prove to be useful to surgeons and radiologists in assessing a patient's status postoperatively.     

Microbiologic and pathologic findings in an epidemic of equine pericarditis.

During the spring and summer of 2001 and in association with the mare reproductive loss syndrome, 22 terminal and 12 clinical cases of equine pericarditis were diagnosed in central Kentucky. Actinobacillus species were the principal isolates from 8 of 10 nontreated, terminally affected and 3 of 10 clinically affected horses. Enterococcus faecalis and Streptococcus zooepidemicus were cultured from the remaining 2 nontreated terminal cases. No viruses were isolated in tissue culture. Nucleic acid of equine herpesvirus-2 was detected in pericardial and tracheal wash fluids of 3 and 1 individuals, respectively. Microscopic alterations in sections of heart and parietal pericardium were consistent with chronic fibrinous bacterial pericarditis. This report confirms a significant role of Actinobacillus species in equine pericarditis and describes an epidemic of this infrequently observed syndrome in the horse.     

Susceptibility of Clematis Varieties and Species to Stem Infection by Phoma Clematidina as an Indicator for Resistance to Wilt

The susceptibility to stem rot and wilt caused by Phoma clematidina was tested for a range of clematis varieties and species widely grown in Europe, and compared with their susceptibility to wilt in practice as judged by amateur and commercial growers in two postal surveys. Most small flowering clematis species did not die back when unwounded stems were inoculated with spores of P. clematidina, and were also reported in the surveys as being very resistant to wilt. In contrast, large flowering hybrids were found to be highly susceptible both to stem infection by P. clematidina and to wilt in practice according to growers. A significant positive linear relationship existed between the susceptibility scores for disease in the stem inoculation test and those for wilt from the surveys. Fungal isolations were made from the stems of naturally wilted clematis plants from British nurseries and private gardens. Unlike the many other fungal species isolated, P. clematidina occurred very frequently, especially in stems of large flowering varieties. Together with the results from the grower surveys and stem inoculation trial, this indicates that this fungus is a regular cause of clematis wilt. The stem inoculation test with P. clematidina described in this study would be useful for European breeders and growers to determine the susceptibility of new clematis varieties to wilt before they are marketed.     

Inheritance of pancreatic acinar atrophy in German Shepherd Dogs

OBJECTIVE: To assess the heritability of pancreatic acinar atrophy (PAA) in German Shepherd Dogs (GSDs) in the United States. ANIMALS: 135 GSDs belonging to 2 multigenerational pedigrees. PROCEDURE: Two multigenerational pedigrees of GSDs with family members with PAA were identified. The clinical history of each GSD enrolled in the study was recorded, and serum samples for canine trypsin-like immunoreactivity (cTLI) analysis were collected from 102 dogs. Dogs with a serum cTLI concentration < or = 2.0 microg/L were considered to have exocrine pancreatic insufficiency (EPI) and were assumed to have PAA. RESULTS: Pedigree I consisted of 59 dogs and pedigree II of 76 dogs. Serum cTLI concentrations were measured in 48 dogs from pedigree I and 54 dogs from pedigree II. A total of 19 dogs (14.1%) were determined to have EPI, 9 in pedigree I (15.3%) and 10 in pedigree II (13.6%). Of the 19 dogs with EPI, 8 were male and 11 were female. CONCLUSIONS AND CLINICAL RELEVANCE: Evaluation of data by complex segregation analysis is strongly suggestive of an autosomal recessive mode of inheritance for EPI in GSDs in the United States.     

Prospective evaluation of laparoscopic-assisted gastropexy in dogs susceptible to gastric dilatation

OBJECTIVE: To determine long-term outcome associated with laparoscopic-assisted gastropexy in prevention of gastric dilatation-volvulus (GDV) in susceptible dogs and to evaluate use of laparoscopy to correct GDV. DESIGN: Prospective study. ANIMALS: 25 client-owned large-breed dogs. PROCEDURE: 23 dogs susceptible to GDV were referred as candidates for elective gastropexy. These dogs had a history of treatment for gastric dilatation, clinical signs of gastric dilatation, or family members with gastric dilatation. Laparoscopic-assisted gastropexy was performed. One year after surgery, abdominal ultrasonography was performed to evaluate the attachment of the stomach to the abdominal wall. Two dogs with GDV were also treated with laparoscopic-assisted derotation of the stomach and gastropexy. RESULTS: None of the dogs developed GDV during the year after gastropexy, and all 20 dogs examined ultrasonographically had an intact attachment. Another dog was euthanatized at 11.5 months for unrelated problems. Two dogs with GDV successfully underwent laparoscopic-assisted gastropexy after the stomach was repositioned. CONCLUSIONS AND CLINICAL RELEVANCE: Laparoscopic-assisted gastropexy resulted in a persisting attachment between the stomach and abdominal wall, an absence of GDV development, and few complications. Dogs with a high probability for development of GDV should be considered candidates for minimally invasive gastropexy. Carefully selected dogs with GDV can be treated laparoscopically.     

Retrospective evaluation of partial parenteral nutrition in dogs and cats

The purpose of this retrospective study was to evaluate the use of partial parenteral nutrition (PPN) in dogs and cats. The medical records of all dogs and cats receiving PPN between 1994 and 1999 were reviewed to determine signalment, reasons for use of PPN, duration of PPN administration, duration of hospitalization, complications, and mortality. Complications were classified as metabolic, mechanical, or septic. One hundred twenty-seven animals (80 dogs and 47 cats) were included in the study, accounting for 443 patient days of PPN. The most common underlying diseases were pancreatitis (n = 41), gastrointestinal disease (n = 33), and hepatic disease (n = 23). Median time of hospitalization before initiation of PPN was 2.8 days (range, 0.2-10.7 days). Median duration of PPN administration was 3.0 days (range, 0.3-8.8 days). Median duration of hospitalization was 7 days (range, 2-20 days). In the 127 animals receiving PPN, 72 complications occurred. These included metabolic (n = 43), mechanical (n = 25), and septic (n = 4) complications. The most common metabolic complication was hyperglycemia (n = 19), followed by lipemia (n = 17) and hyperbilirubinemia (n = 6). Most complications were mild and did not require discontinuation of PPN. Ninety-three (73.2%) of the 127 patients were discharged. All 4 animals with septic complications were discharged from the hospital. The presence, type, and number of complications did not impact the duration of hospitalization or outcome. However, animals that received supplemental enteral nutrition survived more often than those receiving PPN exclusively. Although PPN seems to be a relatively safe method of providing nutritional support, future studies are warranted to determine its efficacy.     

Thoracic radiography,bronchoalveolar lavage cytopathology,and pulmonary parenchymal histopathology: a comparison of diagnostic results in 11 cats

The purpose of this study was to compare the diagnostic results and value of thoracic radiography, bronchoalveolar lavage (BAL) fluid cytopathology, and lung histopathology in 11 cats with spontaneous respiratory disease in which radiography and cytopathology were inadequate in establishing a definitive diagnosis. In these cats, radiographic patterns were characterized as bronchial (n=6), interstitial (n=3), and alveolar (n=2); other features included hyperinflation (n=3), bronchiectasis (n=2), pleural fissure lines (n=2), pulmonary nodules (n=2), atelectasis (n=1), and a tracheal mass (n=1). Bronchoalveolar lavage fluid was unremarkable in two cats. Abnormal BAL fluid showed inflammation (n=5), hemorrhage (n=2), epithelial hyperplasia (n=1), or was suspicious for neoplasia (n=1). Histopathological evaluation revealed inflammation (n=8), neoplasia (n=2), and vascular congestion (n=1). The predominant radiographic location of disease correlated with the same histopathological location in seven cats, and the cytopathological class of BAL fluid was consistent with the histopathological class of disease in seven cats. There was poor correlation between the types of cells found in the BAL fluid and the pathologist's prediction of the types of cells likely to be found in the BAL fluid based on the amount and type of airway cellularity seen on histopathological examination. The results of this study suggest that in some cats, BAL fluid cytopathology does not always correlate with the type of pulmonary disease identified on histopathology. In respiratory diseases where radiography and cytopathology fail to provide a definitive diagnosis, histopathological examination of the lung may be necessary.